AI is a magical typewriter for content marketers …or is it?

How and when to use AI in content marketing remains a huge question. The debate rages on. My friends who are purists see no place for AI in content (where’s the authentic voice?), and my friends who are big on AI adoption see no place without AI (why not use this helpful new tool?).

Personally, I don’t use AI to write. This sentence is being written by hand. That’s my preference, and it’s what works best for me.

In fact, most marketers don’t write entire articles with AI. According to our annual survey, just 11% of content marketers draft articles using AI. Most use it for ideas and editing.

Beyond writing, these are wonderful ways to use AI for content strategy, including ideation, audits and analysis. Today we’ll review five of these and share the prompts that power them. Here’s what we’ll cover:

1. AI for auditing your blog’s calls to action (maximize list growth)
2. AI for topic research (focus on the three best formats for content)
3. AI for initial edits (alignment with content best practices)
4. AI for finding partners and places to publish (PR and outreach ideas)
5. AI for performance analysis (find topic/channel fit)

Beyond CTAs, topics, editing and analysis, there are many other aspects of a content strategy. We’ll touch on those at the end.

1. Blog main page quick audit

Before hitting publish on your next article, take a look at the place where you’re publishing. This is the home of your content marketing program. Does it include all of the key success factors?

• Specific, descriptive email signup call to action
• Categories are aligned with topics (“content strategy”) rather than formats (“videos”)
• Proof points that build the credibility of the content program

The email signup CTA is the most important element on a blog’s main page. And you’ll find bad examples all over the web. Ideally, your signup box is high on the page and aligns with email signup form best practices. Compare:

With this perspective, you may see huge, immediate improvement opportunities. If the issues aren’t obvious, try this audit prompt. The inputs are a link to the page (or full-page screenshot) and a few details about the visitor and your proof points.

Blog Homepage Subscription Audit Prompt

You are an email growth strategist auditing this blog homepage. A good blog homepage does one thing above all else: converts interested visitors into subscribers. Your job is to assess how well this page does that — and make a clear recommendation.

Topic scan — Extract the 3–5 main topics this blog covers from the screenshot and URL. Then answer:

• Are those topics reflected in the CTA copy, or does the signup feel generic?
• Is the blog organized around topics (what you’ll learn) or formats (videos, webinars, guides)? Format-based organization weakens conversion — it tells visitors nothing about whether the content is relevant to them.

Label anything uncertain as an assumption. Then score the page across these eight criteria. The first four are directly on the conversion path and carry the most weight. Present results as a table with four columns: Criterion | Score ( 1–2 / 3 / 4–5) | What you observed | Specific fix.

Conversion-critical:

• CTA clarity and specificity — does the signup feel like a named, valuable product? Do specific topics appear in the headline, description, or button copy?
• CTA proof and trust — is there a subscriber count, testimonial, or credibility signal adjacent to the form?
• Topic-to-CTA alignment — does the CTA promise content that matches what the blog visibly covers?
• Topic-first organization — are categories and featured content organized around subjects, not delivery formats?

Supporting:

• Top-of-page clarity — is it immediately clear what the blog covers, who it’s for, and what to do next?
• Audience alignment — do the topics, tone, and CTA feel relevant to the target visitor?
• Skimmability — can a fast visitor grasp the value of subscribing without reading every word?
• Visual support — does the design make the CTA stand out, or does it disappear into the page?

Strategic gaps — Identify one or two issues the table can’t capture: structural or strategic problems that row-by-row fixes wouldn’t solve.

Verdict

• Overall strength: High / Moderate / Low
• Recommended action: Republish as-is / Light update / Significant rewrite / Rebuild
• Single highest-leverage edit: the one change most likely to improve subscription rate
• Conversion read: would a first-time visitor who cares about these topics feel compelled to subscribe — or leave without a reason to return?

[Upload a full-page screenshot. Optional: note the target audience job title and any proof points: subscriber count, testimonials, or trust signals]

The report rates the extent to which your blog main page aligns with content marketing best practices and email list growth. 

I ran the Orbit Media blog main page through this mini-audit and it definitely had some ideas. It called the way we show our posts a “chronological tile dump” and I can’t disagree.

2. AI for topic research

There are three types of content that tend to outperform all others:

1. Original research (new data)
2. Strong opinion (thought leadership)
3. Topics that aren’t yet covered (new angles)

The lowest performing content is often general, how-to articles that offer no unique methods, data or perspectives. If it could have been written by any brand (or an AI) then it isn’t going to be memorable, be bookmarked, or start a conversation. Maybe don’t write it.

AI can absolutely help you uncover relevant content ideas for strong opinions, original research and hidden gaps. Here are three prompts that do exactly that:

The “Provocative-But-Mundane Topics” Prompt:

What are some relatively mundane, almost trivial topics that [job title] has very strong opinions about?

Everyone should try this prompt at least once. It elicits topics that are triggering without being overly controversial. Here I put in “content strategist” as the job title…

As you scan the response, you’ll feel an almost physical reaction to the topics. You’ll laugh. You’ll groan. You’ll wonder if you could dip your toes into the pool of opinion-based content. Or perhaps dive directly into the depths of true thought leadership.

The magic here is that these are topics that AI can’t touch. No LLM can write strong opinion content. That’s because AI has no opinions. It has no point of view. But you do. Thought leadership is a human-only format.

The challenge is courage: Are you willing to take a stand? Is everyone on the team ok with this? Does it fit within your brand guidelines? Remember, it’s not thought leadership unless someone can disagree with it.

The “Original Research Topics” Prompt

Target audience job title: [job title]

Suggest original research topics relevant to the industry of the role above. The goal is to identify frequently asserted but rarely supported claims — statements everyone repeats but nobody has actually measured. The best topics, once researched, would be cited by journalists, editors, and bloggers looking for fresh data.

Answer these three questions:

1. Where are the gaps? Which topics in this industry are widely discussed but lack solid supporting data?
2. What’s worth measuring? Which unverified claims, if proven or disproven, would be most useful to journalists looking for fresh statistics to cite?
3. What’s the angle? For the two or three strongest opportunities, suggest a research approach — survey, analysis, experiment — and explain what would make the finding credible and linkable.

New research gets cited in other people’s content. It’s catnip for journalists and editors. This leads to links that drive search rankings for both your content and your key conversion pages. These rankings drive traffic through traditional SEO, but also inform AI responses, because of course, AI searches using search engines.

So if you can find a “missing stat” or a common assertion that is rarely supported with data, and then produce that data with a new report, you’ll have plugged a hole in the internet. Go beyond the typical industry trends. Promote it well with other content creators (ideally, by including them in the report as contributors) and you’ll attract links.

I ran this prompt using the “Content Strategist” job title. It had several interesting ideas, along with possible methodologies and angles. None of these would produce serious peer-review level research papers. Any of these would produce interesting, link-worthy content.

The challenge is time: Can you design a study that you can produce without a huge investment? Do you have any data within your company you can use? Or partner with another content program?

The “Topics-Rarely-Covered” Prompt

What are the most important topics in the [job title]’s industry that are the least likely to be covered by the popular blogs?

It’s not possible for you, a human, to read every article in an industry and spot the topics that no one is writing about. But AI does this well. Drop that prompt into any of the AI tools and you’ll see a wide-open ocean of highly differentiated content opportunities. Anything you could cover?

I also ran this one using “content strategy,” and it found some off the less common topics on content marketing blogs. I can easily imagine high clickthrough rates for these…

The challenge is alignment: Which of these rarely-covered topics aligns with your buyer and the rest of the buying committee? Just because it’s new doesn’t mean it fits. 

It’s also possible to combine these methods with good brainstorming, human creativity and subsequent prompts. Imagine if you published a research piece that gave you an opportunity to express a strong point of view. Plant your flag on a solid ground of new data. It’ll be your strongest piece of the year.

3. Article audit prompt

The email signup box is polished. The topics are set. The article is written and it’s ready to go live.

Again, before you hit publish, take a minute to check the new piece against content best practices. In our experience, these are some of the most important elements in any article:

• Strong opening hook: Does the first sentence make them want to read on?
• Evidence and examples: Are the claims backed with data, quotes or specifics?
• Expert or contributor input: Does any voice other than the brand’s appear?
• Formatting and scannability: Headers, bullets, bold, paragraph length
• Visual support: Is there something visually interesting at every scroll depth?
• Internal links: Do they guide readers to related content? Does it link to a service page?
• Personal angle or opinion: Does the article take a clear position, or does it just cover the topic?
• Call to action/next step: Is there a logical next move for the reader?

No, you don’t need to include them all. But you should consider them all. Scan through the final draft and look for content gaps. Better yet, give AI the draft along with this Article Audit Prompt:

Article Audit Prompt

You are a content strategist auditing this article on behalf of an editor who needs to decide whether to update, rewrite, or remove it. Your job is to make a clear recommendation, not hedge.

Score the article across these eight criteria. Present results as a table with four columns: Criterion | Score ( 1–2 / 3 / 4–5) | What you observed | Specific fix.

• Opening hook — does the first screen earn a scroll?
• Evidence and examples — are claims backed with data, quotes, or specifics?
• Expert or contributor input — does any voice other than the brand’s appear?
• Formatting and scannability — headers, bullets, bold, paragraph length
• Visual support — do images or charts add meaning, or just break up text?
• Internal links — are there contextual links to related content? Does at least one link point to a company service page?
• Personal angle or opinion — does the article take a clear position, or does it just cover the topic? A distinct point of view is one of the clearest signals that this content wasn’t written by AI.
• Call to action / next step — is there a logical next move for the reader?

Strategic gaps — After scoring, identify one or two things the table can’t capture: positioning weaknesses, missing angles, or structural issues that no amount of polishing would fix.

Verdict

• Overall strength: High / Moderate / Low
• Recommended action: Republish as-is / Light update / Significant rewrite / Remove
• Single highest-leverage edit: one specific change that would move the needle most

[Upload a full-page screenshot of the article — include top, mid-page, and bottom if the article is long]

The audit pushes you to challenge yourself, to strengthen your assertions, back up your claims and pull in the voice of experts. I ran this on a draft of this article, and it made some strong suggestions. Looks like I have some work to do before this meets the standards of our own content strategy.

4. Discover where your audience gets their content

It’s a mistake to only publish content on your own website. The best content strategists will write for any platform where their audience is paying attention. They network with editors, reach out with little pitches and collaborate with influencers.

If we know where our audience spends time, we know where to listen and where to engage. With the right audience insights, we’ll know the publications to pitch to and the events we can attend. We can meet them where they are.

Can AI help you find these people and places? Yes.

For this, I’ll use a simplified version of Liza Adams “Watering Holes” prompt.

A Simplified “Watering Holes” Prompt

You are a content strategist familiar with [job title and industry]. Identify the “watering holes” where this audience learns, engages, and connects.

For each category, name specific real examples — actual communities, publications, associations, events, podcasts, YouTube channels, forums, newsletters, and influencers.

Create a table with watering hole categories as rows and specific named examples as cells. For each, note the opportunity type: places to publish or pitch guest content, podcasts or channels to pitch yourself as a guest, events to attend or sponsor, communities to listen and engage in, or partnerships to pursue.

Finally, add a short section recommending the three highest-priority watering holes for someone just starting to build presence in this market, with a one-sentence rationale for each.

ProTip! Sign up for the audience intelligence tool, Sparktoro. The free version will show you relevant social networks and websites. The paid version also lists podcasts, YouTube channels and SubReddits. Either way, download the report as a PDF and upload it with the same prompt. 

Here I ran the prompt for “Marketing Directors” in “Healthcare Tech.” I’ve trimmed the results because the original response was a huge list of opportunities.

Scan through and your head will fill with ideas. Here are just a few to get you started:

• Which topics would get traction in which places?
• Look at the past sessions at those conferences. Have we covered those?
• Check out the editors of those publications. Do we know them yet?
• Who sponsors those tradeshows. Anyone we can partner with?
• Listen to a few of the shows. Which influencers does our audience trust?

Even if you’re starting from zero, with no knowledge of this vertical whatsoever, you can quickly see where to start listening. Within an hour of AI-supported research, you may have some great digital marketing ideas for PR, event participation and influencer marketing.

---

Liza Adams, AI Advisor & GTM Strategist, GrowthPath Partners “Most of us use AI to create more content. But it can also help us find where our audience already pays attention. That’s what the watering holes prompt does. It shows you the communities, publications, and events your buyers already trust, so you can go meet them there instead of waiting for them to find you.”

---

5. Give AI your GA4 reports… and align topics with channels

Every content strategist knows how important it is to align topics with marketing channels. Some topics work better in some channels.

Just think about search vs. social. Users of search engines are busy; users of social media are bored. That’s why some topics have opportunities in SEO (detailed long-form content that answers questions) and others do great in social media networks (surprising and visual).

Here we’ll give AI data from our Google Analytics account, showing which articles (title tags) performed how well (traffic, engagement, and conversion rates) across which channels (session source). The analysis is all done by the AI of your choice.

First, we’ll get the GA4 report. This is really just one of many ways you can have AI help with data analysis. First, a screenshot that shows all of the settings.

1. Go to the “Page path + query string” report with a nice long date range.
2. Set the drop-down above the first column to “Page title and screen class”
3. Click the blue plus (+) which is also at the top of that column to add another dimension: “Session source”
4. Set a filter to show only blog posts (“Page path and screen class” contains “blog”) and to remove direct traffic (session source does not contain “direct”) because that isn’t helpful here.
5. Export to CSV.
6. Finally, open the file and remove rows with super low traffic, irrelevant data such as translated titles (non-English) or anything else that looks weird.
7. Save it then upload that to your favorite AI with this analysis prompt

The “Top Channel Alignment” Prompt:

You are a content strategist skilled at using data to identify which topics perform well in which channels. I’m giving you a GA4 report showing blog content performance by traffic source, for a blog targeting [job title]. Suggest 10 new article topics matched to the channels where they’re likely to perform well. For each, note the visitor mindset (e.g., answer-seeking, scroll-stopping), explain the channel fit, and suggest one promotion tactic.

[upload GA4 report]

Like many AI methods, the input is the secret sauce. This one analyzes the performance of your own content with your own audience based on the GA4 report. This isn’t based on best practices. It’s based on what’s worked for you in the past.

Here’s what it looks like from our Analytics. I trimmed it a bit because it was long…

The topics are closely related to your existing content, they’re aligned with channels. For some, you can optimize the content for search. For others, the social media posts write themselves.

The AI even gives recommendations for content promotion. This doesn’t mean that one piece can’t be promoted in all channels. But it helps the content strategist think ahead to distribution, even before the content creation.

AI gave us a boost, but there’s a lot more to content strategy

We’ve reviewed five use cases for AI content strategy, but the ideas here were just the foundation. The target and topics are the beginning, but there are many other aspects to content planning and your content creation process.

• Alignment with business objectives (sales support content)
• Content formats: visual content (videos, infographics), written content (articles, interviews, guides)
• Promotion channels
• Publishing frequency
• Keyword research and SEO strategy
• Technology and tools (content personalization and predictive analytics)
• Team, partners and workflow
• Success metrics

Imagine how ChatGPT or Claude could help with these other aspects of content marketing strategy.
Imagine how the related tasks can get easier, faster and more comprehensive.
Imagine the prompts that would provide the most valuable insights and support your content marketing efforts.

But even with the cleverest prompts and the most interesting AI responses, you’ll need to stay involved. Watch it closely. Stay critical. None of those prompts will let you abdicate your job defending quality, filtering out the noise, adding your own perspective, connecting with the partners and ultimately, the prospects.

We’ve made the case that AI-driven content strategy is better than AI generated content. Generative AI can support the content creator with content ideas.

Then you, the expert, provides the human oversight and decides how to apply the insights.

Stop writing prompts by hand.

That’s my top tip for AI based on hundreds of hours of experimentation and testing. It’s the first of five tips in this guide. By the time you reach the end, you’ll have built an AI assistant that writes your prompts for you.

This isn’t about little ad-hoc prompts. We all use quick little prompts to get help with little tasks in the moment. Those are mostly just for getting answers or getting unstuck.

This isn’t another prompting framework. There are tons of those already. They’re usually acronyms: BRIEF, CLEAR, PRISM, RACE, TRACE, etc. These are really just things to include in your prompts.

Today we’re focused on creating powerful, workhorse prompts that you save and use many times. They may generate reports, write content briefs, edit articles, etc. The kind of prompts that become part of a regular workflow and maybe even part of your daily job.

Here are a few tips for writing those high-performing prompts.

1. Don’t write prompts yourself. Have AI write them for you.
2. Start by talking to AI about your idea. Don’t just tell AI to write you a prompt.
3. Ask AI to challenge your thinking. Get it to push creativity and challenge your assumptions.
4. Create prompts that are fast and efficient. No need to output heavy file formats.
5. Make easy-to-reuse prompts. Then they become part of your workflow.

The best prompt writers use these approaches. They have moved up the AI proficiency curve, from doing things by hand, to getting help from AI, to finally, building AI tools.

You can move up that curve much faster than I did. You can jump right to the end and make a Prompt-Writing Assistant that makes better prompts every time. It incorporates all of the five tips above.

How to make a “Prompt-Writing Assistant”

If you haven’t created an AI assistant yet, today could be your big day.

An AI assistant (sometimes called an automation or agent) is a reusable little tool that does a specific little job. They’re easy to create and easy to share with your team.

Every paid version of AI has this feature. They’re called Claude Projects, Custom GPTs, Gemini Gems or Copilot Agents. They’re all really the same idea: a set of instructions (like a big prompt) that the AI references every time you use it.

We’ve made assistants that research brands, build personas, audit webpages and edit articles. But one of our most powerful is an AI assistant that helps write high-performing prompts. 

If AI is the master of language, then it’s also the master of the language of prompting.

Here are the instructions for this assistant. We mostly use Claude so it’s a “Project” but this will work as a custom GPT or Gemini Gem just the same.

Just open your AI tool, click the button to create, and copy/paste these instructions into the box. If you’re not sure where to click, you can see our guide (and video) on building custom GPTs. It’s literally five clicks. I counted.

In a hurry? We made this a custom GPT that you use anytime: Orbit’s Prompt-Writing Assistant GPT.

Instructions for a Prompt Collaborator (Claude Project, Custom GPT or Gemini Gem)

You help the user design, refine, and publish AI prompts — primarily for marketing and adjacent business functions. Your job is not just to write prompts. It’s to help the user think more creatively about what a prompt could do, what inputs would make it stronger, and how to share the result.

Scope: This project builds workhorse prompts — reusable, high-performance prompts that become part of a regular workflow. Not ad-hoc questions (“summarize this email”), not generic framework templates (BRIEF, CLEAR, PRISM). Workhorse prompts earn their place in someone’s job by producing consistent, decision-grade output across many runs. If a user’s request is clearly ad-hoc, answer it directly without invoking the full methodology. The principles below apply when the prompt will be reused.

How to start a conversation: When the user opens a new conversation about building a prompt, don’t draft immediately. Briefly frame what you’re about to do, then ask 2–3 targeted discovery questions before writing anything.

A good opening looks like: “Before we draft, I want to understand the idea. A few quick questions: 1. What problem are you solving, or what decision should the output drive? 2. What inputs do you have available (URLs, screenshots, transcripts, exports)? 3. What’s the conventional approach here and what part of it usually feels generic or formulaic?” Adapt the questions to the request. The third question should always probe the framing or the conventional approach.

Skip discovery when the user is iterating on an existing prompt, if they’ve front-loaded enough detail to draft confidently, or if they’re asking a quick question that doesn’t require a full prompt. When skipping, draft directly but still apply the principles below.

Principles: These are moves available to you, not steps to execute. Use the ones that genuinely improve the prompt for the user’s situation. Reaching for clever moves (framing challenges, inversions, cross-discipline borrowing) when the straightforward approach is fine produces worse output, not better. It adds friction and makes the project feel performative rather than useful. Inversion, framing challenges, and cross-discipline borrowing should surface as observations and questions about their idea, not as labeled techniques.

1. Push for creative, unexpected angles when the conventional approach feels stale. Workhorse prompts produce better output when their methodology is non-obvious. If the user’s idea sounds like a standard workflow, look for the move that would make the prompt distinctive: reframing the reader’s role, starting from an unexpected input, inverting the process, asking for output the user didn’t think to request. Not every prompt needs this. Reach for it when the user is searching for novelty or when the conventional approach would produce average results. Start with the quantitative data the user likely has access to: open rates, CTR, conversion data, GA4 exports, win/loss outcomes. This is the input most likely to sharpen the prompt. Then suggest non-obvious inputs they may not have considered: transcripts, support tickets, replies, internal docs. The best prompts triangulate both.

2. Match the input format to the question. A webpage can be given to AI three ways, and each surfaces different things: 1. URL or pasted text → copy and messaging 2. Full-page screenshot → design, layout, and images 3. Uploaded HTML → code, schema, title tags, meta tags …Apply the same thinking to other input types.

3. Push for surprise, in both methodology and output. Workhorse prompts produce better results when they go beyond the average, expected answer. Two places to engineer for that:

• In how the prompt is built: look for non-obvious angles — changing the sequence, reframing the reader’s role, starting from an unexpected input.

• In what the prompt instructs AI to produce: build in moves that push past AI’s defaults. Challenge widely accepted beliefs. Highlight common assertions not supported by data. Surface counter-narratives. Reveal overlooked areas of importance. Suggest counter-intuitive approaches. Connect seemingly unrelated insights. Draw from unconventional sources.

Pick the moves that fit the prompt. A trends piece benefits from counter-narratives. A strategy doc benefits from counter-intuitive approaches. A research prompt benefits from connecting unrelated insights. Not every prompt needs surprise. Operational and categorization prompts often shouldn’t have it. Reach for it when the user wants novelty or when conventional output would be too easy to ignore.

4. Challenge the framing only when there’s a materially better one. If the user’s framing is workable, build within it. Only challenge the framing when there’s a clearly better alternative that would meaningfully change the output — not when there’s a defensible-but-different way to slice the problem. Polished pedantry is worse than a polished answer to the question as asked.

5. Look for methods from adjacent disciplines. Frameworks from content strategy, SEO, UX research, paid media, positioning, strategy, or sales methodology can sometimes be repurposed in novel ways. Schema.org as a content brainstorming tool is one example. Worth surfacing when the user is searching for a new angle.

6. When describing what “good” looks like, show an example instead. Examples beat adjectives. A single before/after pair teaches AI more than three paragraphs describing the standard. If the user is describing desired output in adjectives, ask whether they have an example to include instead.

7. When assigning AI a role, include a constraint and say what to avoid. “Act as a world-class copywriter” does little. What works: role expertise plus audience plus constraint. “You are reviewing this page on behalf of a CFO who has 90 seconds, is skeptical of marketing claims, and needs to justify a purchase to a board.” The constraint is where the leverage is. Also include what to avoid, not just what to aim for Example: “allergic to the word ‘unlock’” is more useful than “professional tone.”

8. When a prompt evaluates or compares, require a committed judgment plus rationale. A 0–5 score on specificity, a ranking by impact, a pass/fail on message match, paired with a short rationale. Use a table format. The judgment forces AI to take a position the user has to engage with. The rationale is where the actionable insight lives. Avoid vague feedback.

9. For high-stakes outputs, build in a self-critique step. “After generating, rate this against the criteria above. What’s the weakest part? Rewrite that part.” AI is often better at evaluating its work than producing it on the first try. Make recommendations from any field that may provide insights. Not necessary for every prompt — most useful when the output will be published or acted on directly.

10. When asked to generate a prompt that outputs files, suggest lighter output formats when possible. When a prompt’s output is going to DOCX or PDF, suggest generating a lightweight format like Markdown or HTML first instead. Reserve heavy formats for when the formatting itself is the deliverable.

Of course, I used my own prompt assistant to create these instructions. We worked hard on it together. If you read it closely, you’ll see how it works. It creates a Prompt-Writing Assistant that does a few very clever things:

• It starts by asking questions (so goals are clear and opportunities are discovered)
• It challenges assumptions, crosses disciplines and inverts thinking (so the user gets better responses)
• It thinks about smart inputs for prompts (so the user gives it everything it needs)
• It builds prompts that include little self-evaluations and rating systems and rationale (so the user thinks through the response)
• It builds prompts that generate efficient output formats (so the user doesn’t use more credits or carbon than necessary)

You’ve pasted in those instructions to your Claude Project (or Custom GPT or Gemini Gem) and given it a name. A new member of your team is born.

Run a test prompt to check the quality

We’ve tested this carefully across many use cases and AI models for hours. But you’ll want to test it yourself.

Ready for your first chat with your new Prompt-Writing Assistant? Here are a few conversation starters you can try as test prompts. The use cases are infinite.

• “Write me a prompt that writes email subject lines and body text when given a draft article.”
• “Write me a prompt that generates content briefs when given a topic.”
• “Write me a prompt that edits articles for brand voice and tone.”
• “Write me a prompt that helps me improve underperforming landing pages.”
• “Write me a prompt that writes case study intros.”
• “Write me a prompt that creates internal linking strategies for an article.”
• “Write me a prompt that audits a competitor’s homepage for messaging weaknesses.”

Try one of these. The assistant will draft you a fresh prompt. Review it first or copy it straight into a new conversation to test the output. Here’s what a test looks like.

I used the last example from the list above (audit a competitor’s homepage) and generated two prompts: one with the Prompt-Writing Assistant, one with a plain unguided request to the same AI. Then I ran both side by side. Compare:

In this test, the difference is dramatic. The Prompt-Writing Assistant generated a prompt that had much more detail and insights.

• It started by asking me questions (“What decision are you driving toward?”)
• It suggested inputs (“Can you provide a screenshot?”)
• It created a scoring system (“Ten-second test results: 2.5 out of 5”)
• It critiqued its own output (“The weakest judgment in this audit was scoring the hero based on a single screenshot of a rotating carousel.”)

Of course, they are both better than a one-shot prompt written by hand “Audit this homepage for messaging weaknesses.” I tried that too. It was a list of generic recommendations. Useless.

What do you do next?

1. The output is good…
   Save your new prompt! (more on this below)
2. The output is not great…
   Fix the issues in the new prompt. Edit it then fire the test prompt again.
3. The output is way off…
   There’s something structurally wrong, useless or irrelevant. Training time. Go back and tell your prompt assistant what’s not right. Ask it to create a revision to its own instructions. Then copy the revised instructions, paste them in, start a new conversation and fire the test prompt again.

This more diligent approach to dialing in your Prompt-Writing Assistant involves checking it in several stages.

It’s a strange and abstract way to work. You’re using an AI assistant to generate new prompts that you’re testing with test prompts in another conversation, sometimes with the assistant auditing its own output. Layers of language. It’s mind bending work with multiple levels of editing. It’s like that Leonardo DiCaprio movie ‘Inception’. Great movie.

Keep iterating until you have an assistant you trust and a final prompt worth sharing. This approach is how we’ve made many of our most effective prompts.

Use your Prompt-Writing Assistant to improve your old prompts

You may already have a little repository of prompts you love. They’re very good. But could they be better? Could the analysis prompts provide even deeper insights? Could the content prompts generate even more complete articles?

There are two ways to use your new Prompt-Writing Assistant to audit and improve them

1. Give your original prompt directly to the Prompt-Writing Assistant
   “Here’s a prompt I’ve been running regularly. Audit it against the principles in your instructions. Suggest specific improvements that would make it sharper, more reusable, or more decision-driving. Don’t change its purpose or the shape of its output. Tell me which principles each suggested change is grounded in.”
2. Have the Prompt Writing-Assistant make a new prompt, then give both prompts back to the assistant
   “I’m giving you two prompts that solve the same problem: one you generated, one I’ve been using. Compare them across the principles in your instructions. Where does each one outperform the other? Then write a third prompt that combines the strongest elements of both.”
3. Run both prompts and compare the actual outputs.
   “I ran both prompts on the same input. Here’s what each one produced. Which output is stronger, and why? What does the weaker prompt need to match or beat the stronger one?”

The same approach works to audit the instructions behind Custom GPTs, Claude Projects, Gemini Gems, and Claude Skills. Paste in the instructions, ask the Prompt-Writing Assistant to audit them against its principles, and you may spot gaps that have been quietly limiting every conversation you’ve had with that assistant.

---

Andrej Karpathy, OpenAI Cofounder “The hottest new programming language is English.”

---

Give your new reusable prompt a home

A workhorse prompt without a home gets rewritten from scratch by the next person who needs it. Or by you, three weeks from now, when you can’t remember where you saved it. Put your workhorse prompt in the stable. This could be several different places:

• Save it in a shared prompt library that your team can easily access
  This could be Google Drive, Sharepoint, Notion or whatever else you use to share docs. Then link to it from your team training docs and SOPs (“standard operating procedures”)
• Make the new prompt its own custom GPT, Claude Project or Gemini Gem
  If it’s used all the time, no need to copy and paste it every time. Just make it another new assistant. Give it a name “Orbit Email Editor” or “Cletus the Copywriting Helper.” As long as you have a team AI account, everyone can access it.
• Claude Users Only: Make it a “Skill”
  Tell your Prompt-Writing Assistant to adapt it and make it a Claude Skill. Skills get invoked automatically when the request matches the request. The user doesn’t need to go to the assistant.

Make to your Assistant ever-smarter (upload knowledge sources)

What would make your assistant smarter? What could you upload to it that it could reference when generative prompts?

That depends on its purpose. If it’s generating prompts that do editing, upload some writing standards that it can incorporate. If it’s generating prompts that make reports for clients, upload your company style guides or a past report. If it’s generating prompts that do audits, upload your current methodologies.

Our page auditor has several interesting knowledge sources that it can reference.

Anytime, anywhere. If you spot something that might make any of your AI assistants smarter, save it, convert it to a lightweight markdown (.md) format, and upload it to the knowledge sources section of your assistant. Your assistant may be a bit smarter forever after.

I have many marketer friends who bristle at this kind of AI delegation. They have valid reasons, from labor market impact to environmental costs. But one of the strongest objections is just about the future of work and its intrinsic value to human purpose. What about the simple joy of making things? What about that?

The joy of creativity is alive and well

In December of 1999, I quit my recruiting job to become a web designer. I wanted to build websites more than anything else in the world. By 2001, my cofounder and I had started Orbit Media as a web development company and we never looked back. I haven’t personally designed a website in many years, but I still remember the joy of making things.

And I still feel that creative joy, but it’s moved. I feel another kind of joy, The joy of building little tools and seeing them work. It’s still creative, but it’s a different kind of creativity. Maybe it’s more like the joy that a programmer feels when their code does a good job.

And of course, the greatest joy is performance. Knowing that you made a difference for a brand.

Marketing is important work. It fills the pipeline that drives deals that drive the revenue. That revenue creates job security for the humans that work at those brands.

With this singular purpose, we must consider any tool that drives results. AI is one of them. But the point was never AI or any other tool. The point is the impact that the work has on the business.

And now that we can make our own little tools, we can make one shift that secures the future for all of us.

Focus on outcomes, not outputs.

We are marketers, not artists. We take pride in our work, but we must also take pride in the performance of our work. Because marketing matters.

Every AI prompt has two prices.

One shows up on your invoice. The other gets paid in carbon, electricity and water. Both are driven by token usage and what you upload to and export from the AI. And the file formats you choose can change the bill by 10x.

Why do some files require more processing?
There are three factors that combine to determine how hard the AI has to work with any file.

1. Does the AI need to use a tool to open the file or generate the output?
2. How much structural noise (tags, formatting, styles) does the file format require?
3. Is anything being processed as an image instead of text?

Some file types that have lots of formatting and visuals. These require much more processing power. Here’s a simple way to remember this:

• If you need software to open it, it’s expensive (DOCX, PPTX, XLSX, PDF)
• If you can open it in Notepad, it’s efficient (TXT, MD, CSV, JSON, HTML)

If you need some desktop software to open the file, then the AI needs libraries and tools to parse the file. These libraries and tools add tokens and carbon cost. It may be obvious. You can sit back and watch the AI work harder (and slower) on these files.

Imagine you’re working on a big report and getting help from AI… 

You can generate all kinds of file formats using AI. Which do you choose?

To get a sense for which file formats are the most popular with marketers, we put a little poll on LinkedIn. We had about 500 responses to the question “When you ask AI to make a big deliverable, what format do you ask for?” The top answer was DOCX by a wide margin.

But those are all very different types of files, each requiring very different levels of effort from the AI to generate. They have different costs in terms of tokens (processing power), electricity (carbon output) and credit usage (money).

Which file formats have the highest costs?

To measure that difference, I took a recent article and gave it to Claude in six different file formats, then analyzed the token count using the “v1/messages/count_tokens” API endpoint. I measured in both directions:

• Input Measurement (added to prompts): I asked the AI to “summarize” the piece as provided in the six formats, including a nicely designed PowerPoint version of the article that I’ve used in presentations.
• Output Measurement (generated for downloading): I then reversed the process and asked AI to “generate” the content in those same file formats.

Here is the relative cost in tokens, which translates directly into energy consumption, environmental impact and financial cost. You can see that mileage varies.

The difference is huge. HTML and PDF are expensive to read but cheap to write. DOCX is expensive in both directions. PPTX is by far the heaviest, slowest and most expensive in the bunch.

Let’s look closer and consider more file formats. This table explains how AI sees the files and handles the processing, in and out.

Let’s look closer at the rankings, share tips and consider lighter-weight alternatives to the heavy, popular options.

The formats, use cases and alternatives

1. Text Files (TXT)
The smallest, lightest file possible. But contains no formatting. Good for large amounts of text where formatting isn’t important, such as podcast transcripts.

Lighter alternative: None. 

2. Markdown Files (MD)
Markdown is the language AI agents speak best. It’s also readable by humans and not much heavier than a TXT file. Formatting and tables are added through characters, not tags. It’s the fast, efficient champion of AI inputs. It’s also great as a draft that you can edit within the AI, getting all of the content ready before publishing to a heavier format.

Lighter alternative: None. 

3. Comma Separated Files (CSV)
Lightweight, structured and easy for AI to parse. Great for simple tables, lists, exports, keyword research, analytics data and CRM records. No formatting, formulas or multiple sheets. Just rows and columns.

Lighter alternative: If it’s just a short list or simple table, paste it directly into the prompt.

4. Web Pages (HTML)
HTML was  our least common answer in the LinkedIn Poll above, but they have many advantages. They’re not too heavy to generate and easy to share (everyone has a browser) Maybe the ideal for reports. Then can even include interactivity, but this adds a lot of weight.

HTML files don’t play nice with Google Docs. You can’t paste HTML into a Google Doc and if you upload them to Google Drive, they don’t render the HTML when you open them.

Lighter alternative: If you don’t need print-ready formatting, MD carries the same content at a fraction of the cost.

5. Word Docs (DOCX)
This is the most common file format for outputs, according to our little LinkedIn Poll above. No surprise. It’s still a standard for word processing. But it’s a heavy file format with lots of tags and formatting. It wasn’t built for AI.

Like all archive formats, these need parsing on the way in and libraries on the way out.

• Instead of importing a DOCX file to AI, convert it into an MD file and then upload (use Google Docs or save to TCT then change the file extension)
• Instead of exporting a DOCX file from AI, make an MD file then paste it into a fresh DOCX file (fast, easy, but may require a little manual styling)

Lighter alternative: MD for drafts, HTML for polished-looking documents. 

6. Excel Files (XLSX)
Useful when you need multiple sheets, formulas, filters, formatting or charts. But all of that structure adds weight. For AI inputs, XLSX files are overkill unless you need formulas and logic.

Lighter alternative: If you don’t need formulas or formatting, CSV does the same job at a fraction of the cost.

7. PowerPoint Files (.PPTX)
These have a very high cost output. They require a ton of processing. Every layout decision in every slide is all code that the AI has to read. And images require a separate, slower model. Those visuals drive up the bill.

Better yet, just don’t mix AI and PPTX files if you’re concerned about costs. If you need to make a PPTX in AI, produce the output just once, at the very end of the process.

And don’t use AI for small edits to decks. It’s faster and cheaper to make little changes manually in PowerPoint. If you must edit within the AI, be explicit: “Just regenerate slide 3, keep everything else the same.”

Lighter alternative: If the slide format isn’t the deliverable, an HTML file carries the same content far more cheaply.

8. PDF Files
If you’re building a Custom GPT, Claude Project or Gemini Gem, you may be tempted to upload some relevant PDFs. But before you do, start with this prompt: “I’m planning to add this PDF document to your knowledge sources, so let’s convert it to an MD file first. But before we do, what should be trimmed? Or consolidated”

The AI will suggest edits that will reduce costs without sacrificing quality. And the smaller file (MD, not PDF) will make all future conversations with this project cheaper and lower-emissions. It’s a double-win.

Lighter alternative: Use MD for text-heavy documents, HTML for designed reports

9. Images (PNG / JPG)
AI image generation is not done with an LLM. It’s a separate model that uses “visual tokens.” If you’ve had AI make pictures, you’ve seen how slow it is. That’s because it requires a lot of processing power.

If you’re only worried about the financial costs and not the environment, you can generate images in a cheaper model (Ideogram is the best) and save the credits in your daily workhorse AI, where you sometimes hit credit limits.

Lighter alternative: For screenshots of text, just paste the text instead, if possible. 

Where costs multiply: editing files in AI

Here’s where things get really expensive. It isn’t obvious, but all marketers should know this:

Every iteration on a heavy file format costs as much as making it from scratch. 

That’s right. If you ask for even a small change to that DOCX file, the AI may start over and create the document all over again, with all of that token cost, energy use and carbon impact.

Smart AI users know to work on files in a simple format, such as an .md file, where the costs are much lower. Iterate in text, then generate the final format once everything is final. Here are a few tips for efficient AI prompting:

1. Brainstorm first, no file generation yet. Tell the AI what you want to make, but instruct it not to make the final format yet… “Don’t generate the DOCX output until I ask for it. Let’s refine the idea first…”
2. Ask for lightweight previews. While working on the draft, ask it to show a simple markdown format you can preview… “Show a preview in a lightweight markdown format so I can review it efficiently.”
3. Iterate in the text. Have the conversation about content in MD or right in the chat. Get the structure right, get the wording right, get the order right, all in plain text, where iterations are cheap.
4. Batch your changes. If you have five edits, give all five in one prompt. Each prompt that edits a heavy file pays the full regeneration cost. Five separate change prompts is 5x the cost. One consolidated change prompt that lists five edits is 1x the cost. “Change X to Y and move it up in the doc. Remove Z completely.”
5. Ask for the final format once, in the end. Don’t ask for the output until you’re ready for showtime. This is the single biggest efficiency win. All set? Content locked in? Now have it make the big file once and once only. “Changes are final. Now generate the DOCX file. Use these styles…”
6. One more final tweak? Do it outside the AI. Export the big file and make final final changes (there’s always one more, right?) after it’s landed on your computer. Download it → Review it → Tweak it → Ship it

Having AI make tweaks to a DOCX or PPTX file is slow, processor intensive and uses more energy than necessary. It’s like driving a giant truck a half block to pick up a few groceries.

Let’s ask the experts: “Which are the best file formats for AI?”

Marketing veterans and AI experts make thoughtful cases about AI efficiency and file formats.

---

Chris Penn, Trust Insights Choose the smallest model that will get the job done. “In tools like Gemini, ChatGPT, and Copilot, there are a range of model choices but they all tend to fall in two categories: fast and dumb (instant, fast, Flash, etc.), slow and smart (pro/thinking/etc.) The smaller and faster (and dumber) the model is, the less energy it uses. My rule of thumb: if you’re doing a task that requires less than two rounds of revision, use a fast model. Example: summarizing an email, etc. If you’re doing a task that requires more than two rounds of revision, use a slow model, but know it imposes a much heavier carbon cost.”

---

John Jantsch, Duct Tape Marketing HTML wins on both ends. Cheaper to make, cheaper to change. “I’ve been making slide decks for small business clients for 20 years. Most get opened once and never looked at again. So why am I burning an hour designing in PowerPoint when an HTML page does the same job and renders on any device the client owns? Same logic with SOPs. Nobody wants to crack open a 14-page Word doc to figure out how to post on LinkedIn. They want a link they can scroll on their phone. The reason we didn’t already ship HTML is simple. You had to know HTML, or fight with a website builder. PowerPoint and Word had visual editors that a non-coder could use. AI flipped that. You describe what you want in plain English, and you get a self-contained file you can open locally or host for free. Need changes? Just ask. It rewrites in seconds. It’s also lighter on the planet. HTML is plain text that the AI types directly. DOCX and PPTX are zip archives of XML that the AI has to assemble with separate code libraries. That burns far more processing every time you generate a file. And every time you ask for a small tweak to a DOCX, the AI usually rebuilds the whole document from scratch. The same edit on an HTML file costs almost nothing. Multiply that across a year of client work, and the gap is enormous.”

---

John Nugent, Magnolia Slides are becoming obsolete “The advancements in AI tools in the last 4-5 months have made slide/presentation tools almost obsolete. You still need to deliver in a client-friendly format, but if you’re doing custom HTML right, they won’t want to receive anything else.”

---

Chris Penn, Trust Insights XLSX files are a nightmare “For tabular data, LLMs struggle natively with anything in a table format. Want to represent your data in a machine-readable format that’s also eternally compatible? Ask for JSON or even better, YAML. It converts tables into lists, and LLMs can use them perfectly – AND modify them easily. When you’re done with the LLM, use a free program like jq or yq to convert them back into a tabular format of your choosing, no AI needed.”

---

John Jantsch, Duct Tape Marketing Markdown is the next move. “Drop a free tool like MD Reader into your browser, and any .md file renders cleanly the second you open it. (This is what Wikipedia runs on.) Your team’s playbooks and SOPs look like a real page without anyone fussing over fonts. You write in plain text, and the browser handles the rest. All while using the most token-friendly file that exists. Perhaps today is a new beginning for you: a new format for your big deliverables (HTML) and a new file format for everyday AI collaboration (MD).”

---

So far, we’ve covered things you can control. Now we’ll pan out and see how your prompts fit into a much bigger picture…

The AI cost you can’t control: training the models

Prompting isn’t the biggest cost of AI. It’s training the models. Building a model and keeping it trained requires huge amounts of energy. Every big tech firm with a frontier AI model spends at least $100M per year on training, emitting tens of thousands of tons of carbon (and methane, nitrous oxide, etc.) not to mention water (evaporative cooling to keep servers from overheating).

This is not something users like you and I can control. No matter how much or little you use AI, these companies will continue to invest, incurring financial costs for themselves and environmental costs for us all. Compare the cost for AI model training and user AI prompting:

When asked about this, Sam Altman recently brushed off concerns about AI’s energy use, comparing it to the food humans eat.

“People talk about how much energy it takes to train an AI model, but it also takes a lot of energy to train a human. It takes about 20 years of life, and all the food you consume during that time, before you become smart.”

Sounds dismissive. But he has a point about the high carbon cost for human consumption. So let’s look at that…

If you’re really concerned about your carbon footprint…

There are two main inputs that drive the carbon footprint for you, me and everyone else on the planet: air travel and meat eating. AI aside, any of us can reduce our carbon footprint by cutting back a bit on these activities. This puts AI use in the proper context of each of our true individual impacts.

Shrink your AI carbon footprint

My friends who are the most proficient with AI talk about “token efficiency.” They work hard to reduce credit expenditures. They run the cheapest model possible for each specific job. They put data in tables to avoid prompting completely when possible. They even install AIs locally on their laptops for simpler tasks. The best AI users obsess over efficiency.

But you don’t need to be an AI pro to be efficient. Smart decisions about file formats saves processing power, carbon emissions, water use and yes, money.

Minimizing the impact of AI on the planet is all of our jobs.

Great webpages are great in different ways; bad webpages are bad in the same way. They lack specificity.

A page with details and specifics works hard at all three of its main jobs:

• Provides the key info that converts human visitors (clarity)
• Uses phrases that drive search rankings  (relevance)
• Trains the AIs on the details of your offer (accuracy)

A vague page is less likely to do any of those things well. And it’s less likely to perform in every important way.

Of all the little AI audits you can run on a webpage, this one may be the best. It checks for specificity in five places where vague pages tend to fail:

1. Headers
2. Hero image
3. Subheads
4. Navigation labels
5. Calls to action

We’ll start with the prompt, then we’ll explain the logic behind it in each of the five areas. At the end of this article, I’ll show you a mini-report that you can generate from the prompt’s response.

The Specificity Audit Prompt

This prompt has a special feature. Not only does it check the specifics of the page, but it also checks whether the call to action mismatches the text on the page. This is a type of path analysis, which is easy to miss when you check pages individually.

But to do that, it needs to run in an AI (Claude, ChatGPT, or Gemini) with web access enabled so that the tool can follow the CTA link to the destination page.

The inputs are the full-page screenshot of the page and the URL. That’s all it needs. No need to even upload your persona for this one.

Audit this webpage for vague language in the five elements that most affect clarity: H1 header, hero image, subheads, navigation labels, and CTAs. Provide the URL (and a screenshot if helpful).

For each element, return a markdown table with four columns: Element | Score | Current | Rewrite.

Use this scoring format, where 1 = generic (could apply to any company) and 5 = specific to this business: 1 · 2 · 3 · 4 · 5

The Current column shows what’s on the page now. The Rewrite column shows the exact replacement — not advice about what to change. Flag any word that could apply to any company in any industry (“solutions,” “services,” “learn more,” “get started”). For the hero image, describe what a more specific image would show.

Then follow the primary CTA link to its destination page. Compare the promise made by the CTA to what the visitor actually lands on. Is there a message match, or a gap? If there’s a gap, name it in one sentence.

Prioritize the single highest-impact fix at the end.

[Provide a URL and a full-page screenshot of the page]

The analysis is both enlightening and practical. It includes the problems and the remedies. We’ll show an example report below. But first, let’s look at the five areas we audited, side-by-side with examples…

1. Write Descriptive <h1> Headers

Every page (in fact, every scroll depth on every page) has a visual hierarchy. One element is always the most visually prominent. It attracts the visitor’s attention first. A second element comes next and so their eyes move there next, and so on down the page. That’s your visual hierarchy and it is the key to web design and usability.

We’ll start with the <h1> header because it’s at the top of the visual hierarchy. If it’s generic, vague and boring, that’s bad. If it’s specific and descriptive, that’s good. You’ll pass the 5-second test (more on that below).

The first question of every visitor to every webpage is “Am I in the right place?” The header should answer that question. Compare these two and ask which does a better job of answering that question…

The difference is huge. And it’s an easy opportunity to improve. Making your header more descriptive may take just a few minutes. If internal stakeholders push back, tell them it’s just a test (all changes are, really) and that you’ll change it back if it doesn’t perform well.

Clear is more important than clever in headers. The most helpful, compelling thing should also be the most visually prominent thing. But on websites everywhere, the header is vague but the small text below it is specific. Just flip that and you’ll have a better page for visitors, search engines and AIs.

2. Use a relevant hero image

Also at the top of the visual hierarchy is the featured area image. If the header is specific, then the image doesn’t have to work as hard. The visitor may already know they’re in the right place.

But a relevant image can show what you actually do. But a generic image (often a stock photo) is a missed opportunity to be specific (and authentic). A stock photo doesn’t hurt, but it doesn’t help either. It tastes like water.

Take a look at that image at the top of your page. Ask yourself: “Is this specific to my business, or generic to millions of businesses?”

The hero image doesn’t need to be a picture at all. Sometimes a texture or looping background video. These are fine options that reduce the prominence of that element. The visitor gets clarity from other elements.

Other times, the hero area is the perfect place to show the product (example), show the product in context (example), show the team (example) or show a video where you explain the value of the offer (example).

3. Rewrite (or remove) generic subheads

Once they know they’re in the right place, the visitor starts scrolling and scanning for answers to their questions. That’s why you have <h2> subheads. They are typically the most visually prominent element at every scroll depth. So they should be compelling, specific and descriptive.

Generic subheads just take up space without adding any value at all. They simply push down the rest of the page. Look at your subheads and ask yourself: “Could this be more specific? Would this page be just as helpful if I removed this completely?”

Subheads are an opportunity to put something relevant and helpful into view for the visitor. The best subheads are specific to your business. They communicate quickly, adding value to the experience of the visitor.

Ironically, “What we do” doesn’t actually say what you do. It’s not helpful. “Our solutions” is just as bad. If you ran the prompt above, you already have ideas for more descriptive subheads.

Testimonial subheads are another good example. The subhead “What clients say about us” doesn’t actually show what clients say. Far better to use the subhead to showcase a compelling part of a testimonial.

In the first example above, the visitor isn’t likely to read those super compelling words in the actual testimonial. Again, the most compelling thing is the least visually prominent thing. In the second example, the visitor is very likely to read that part of the testimonial, because the visual hierarchy aligns with the messaging priorities.

---

Talia Wolf, GetUplift “Often, brands stay on the surface level instead of actually going deeper. Social proof is a prime example. We all know it matters, so everyone uses it, but usually that just means slapping a few logos on the page and adding a generic testimonial like ‘they’re the best.’ But no one is sitting there thinking, “I just need to know they’re the best”. They’re thinking: “Will this work for me? What if I get blamed for this? Am I about going to lose my job?” Social proof is your best chance to reduce that risk, so it has to be specific.”

---

4. Use descriptive navigation labels

The navigation labels in your menu are another opportunity to be specific (or super generic and vague). On a homepage, the navigation bar is always near the top of the visual hierarchy. Eye tracking studies always reveal this. Most visitors scan the main menu first, even before scrolling at all.

The goal of your homepage is to get your visitor off the homepage. And of course, the copy is absolutely critical for keyphrase relevance (search rankings) and training LLMs on what you do (AI recommendations).

Specific, descriptive nav labels are part of website navigation best practices. They both tell the visitor what you do and guide them toward deeper pages, help the visitor segment themselves, finding pages more specific to their needs.

Labels such as solutions, services, products and industries are generic, vague and unhelpful. Literally any business could use these.

Scan through your website navigation and ask yourself: Are these labels specific to our business? Or are they common to millions of businesses?

Of course, businesses with many offers may have a good reason to use generic labels at the top level of the navigation. It makes sense. But even huge companies can be very specific in the sub-navigation labels within their mega menus.

5. Craft calls-to-action with compelling verbs

Finally, we get down to the CTA. It’s in your header navigation, partway down your page, and probably, in its own page block above the footer. You’ve got several. But how specific are these CTAs? Are they relevant to your business, or generic to all businesses?

Compare the difference in these two examples. The first is generic and common to every website on the internet. The second is specific to the visitor. It gives them a reason to click.

Look closely at the verbs on your page. Verbs are key. Some calls to action have common verbs like contact, read, learn and click. Great CTAs have specific verbs like schedule, chat, check, download and watch.

For ideas and a helpful prompt, here’s our guide on creating high click-through rate CTAs.

Remember, no one clicks anything until they’ve done a split-second cost/benefit analysis. You can increase the clickthrough rate of anything by making the perceived cost smaller or the benefit bigger. That’s why low-commitment CTAs often work well.

When they click the CTA, the words on the contact page should match the words they just clicked on. Otherwise, the mismatch may cause a bit of friction at exactly the wrong moment. This is one of our best practices for contact pages.

---

Pam Didner, B2B marketer and Copilot Trainer “Good writing is in the details. In a novel, you need specific moments, tension, and turning points to keep people reading. The same applies to business content if you want storytelling to land. Too many blog posts and eBooks stay at the surface. They sound right, but they don’t help. They rely on clean frameworks and tidy steps that look polished but don’t solve real problems. If you want storytelling to work, go deeper. Be specific about features and benefits. Show exactly how something works and why it matters. Walk your audience through the logic, and give them a reason to care – your emotional hook.  Start with why it matters. Then show how it works. Then make it clear what to do next.”

---

The Output: A Specificity Audit Report

Once I got the response from AI, I used this simple prompt to convert the response into a mini-report.

Make this into a simple HTML report.

Why HTML? Because it’s lightweight, portable, easy to edit and universally viewable. You could make a DOCX file or a PPT, but why burn the credits? These are heavy file formats requiring a lot of processing for the AI to generate and edit. Besides, why use PowerPoint unless you absolutely have to?

The report is more visual and interesting than the AI response. Now you’ve got something you can bring to a meeting.

When I create reports, I do it in a little automation (Claude Project, custom GPT, etc.) that knows the Orbit Media styles, so everything aligns with our brand. Here’s part of what it looks like…

Ouch! AI doesn’t like my subheads. It suggests more detailed and specific alternatives. And it has a recommendation for each deficiency it identifies.

As always with AI audits, we should always feel free to dismiss any advice that we find off-target or irrelevant. Better yet, talk to the AI about the report. The audit is really just a point of view, not prescriptive instructions. You, the human marketer, decide what to do next.

The 5 Second Test: Check your revisions with actual humans

You’ve made some changes! The draft looks good. You’re ready to publish …but if it’s a high-stakes page, you may want one final check to confirm that your page communicates quickly and clearly.

The “5 Second Test” is a clever and simple way to test. Upload a screenshot to Lyssna and enter one question: “What does this company do?” The tool then shows the screenshot for five seconds to real humans who then answer the question. You pay $1 per response.

Spend $25 and in a few hours, you’ll see 25 answers.

If the answers look like these, you have a problem…

When we show the results of these tests to clients, everyone suddenly feels urgency to fix the problem. And the fixes add specificity and clarity. Sometimes the fixes are to remove distractions.

So why do marketers use vague language?

If using specific, descriptive language is good for converting humans, ranking in search engines and training the AI models, why is vague language so common on websites?

The main reason is fear.

Marketers are often afraid of excluding one of their audiences. They want to address many different markets. To do so, they water down their language. Ironically, they are so worried they’ll exclude someone that they end up excluding everyone.

Let’s ask Doug. If you don’t know Doug, here’s my standard introduction: Doug Kessler is your favorite writer’s favorite writer. He’s that good. And he has an explanation for vague copy:

---

Doug Kessler, Velocity Partners “The urge to try to be all things to all people is so strong it’s almost the default setting for marketing teams. Excluding a potential buyer (no matter how unlikely it is that they actually do buy) just feels wrong to marketers. And it can be hard to convince stakeholders (especially voracious sales carnivores) that laser-focusing only on ideal prospects is actually a better strategy. The all things to all people mindset leads to fuzzy, flabby, lowest common denominator copy that just doesn’t resonate with anyone.”

---

You invested weeks into that original research piece. You devoted days refining a thought leadership essay, one that encapsulates your brand’s philosophy so eloquently. And yet, you click “publish,” share it on LinkedIn, and watch your traffic soar, only to return to normal 48 hours later.

That’s content decay in a nutshell. And marketers are stuck in this cycle. You create content, you promote content, you abandon. And you’re left with burnout and decreased ROI.

Your brilliant ideas shouldn’t have an expiration date. If your audience is too busy to sit down and consume a 2,000-word whitepaper, they’re not ignoring you. They’re just multi-tasking. They’re commuting, exercising, or walking their dog.

To reach them, you need to transition from screen to earbuds. Audio content isn’t just a new channel.

It’s a new life for your content. Audio content bridges the space between a lead and a listener, a level of intimacy that written content can’t match.

Why Audio is the “Unlock” for Modern Content Strategy

Before we go into the tactics, we must recognize the change in consumer behaviors. Screen fatigue is a very real thing. According to recent industry data, the average professional is looking at a screen for more than seven hours a day.

Source

When a consumer closes their laptop, they’re not closing off their desire for information consumption. Instead, they’re engaging in passive consumption. This is the “eyes-busy, ears-open” consumer market.

• Multitasking capability: Audio is the only media form that doesn’t demand 100% visual attention from the consumer.
• Emotional connection: Tone, sarcasm, empathy, and authority are much easier to convey with the human voice than any font choice.
• Accessibility: Making audio versions of your information increases your brand’s accessibility to consumers with visual impairments or learning differences.

11 Strategic Ways to Repurpose Your Content for the Ear

Repurposing is the ultimate hedge against content decay. It allows you to take the intellectual property you’ve already invested in and deploy it across a channel with significantly higher engagement rates. Here is how to build your audio ecosystem.

1. Convert High-Performing Blogs into Narrated Articles

The easiest way to begin is to offer your visitors the option to listen to your content. Long blog posts are like a goldmine for audio content.

Many people prefer “listening” to a blog post while they’re doing chores instead of scrolling through a mobile device.

You can use tools like ElevenLabs or Play.ht to create high-quality, AI-generated narrations that sound incredibly human.

You can embed this audio player at the top of your blog post. This increases “Time on Page,” which is a crucial metric for SEO.

Source

But don’t just read the text aloud. Create a short, informal introduction to make it sound like a personal message to the reader.

2. Transform Research Data into “Audio Snacks”

If you’ve done original research, then you’ve got a wealth of statistics at your disposal that are great for social media. Instead of doing a static infographic, consider turning your statistics into 60-second audio clips.

You can use a tool like Headliner to create audiograms, which are static images with a moving waveform and text.

These do much better than static images on LinkedIn and X.

• Highlight one unexpected statistic in the clip.
• Share the “why” behind the statistic.
• Share your full report in the show notes or comments.

3. Take Advantage of Captive Audiences with Podcast Advertising

Podcasts have become one of the most trusted media environments for audiences. Listeners often spend 30 minutes or more with a single episode, creating a level of attention that’s difficult to replicate with other digital channels.

For marketers looking to extend the reach of their existing content, podcast advertising offers a powerful entry point into this ecosystem. Instead of launching your own show, you can promote your ideas, reports, or thought leadership within podcasts your audience already listens to.

Audio ads work particularly well for amplifying existing content. For example, a short ad can highlight a key insight from a report, invite listeners to explore a resource, or introduce a new perspective on an industry issue.

Platforms like AudioGo allow marketers to place targeted audio ads across podcasts and streaming platforms, helping brands reach listeners based on demographics, interests, and listening behavior.

By meeting audiences inside the shows they already trust, audio advertising helps extend your content strategy beyond the screen and into the moments when people are commuting, exercising, or multitasking.

Source

4. Turn Case Studies into Audio Testimonials

Case studies are often dry. They follow the “Challenge, Solution, Result” format. Audio changes that by adding emotion.

Ask your client to jump on a 10-minute call on Zoom. Record the audio as they talk about the relief they feel when your solution solved the problem.

• Cut this into small pieces.
• Embed the audio next to the text quote in your case study.
• The human voice creates a trust factor that text does not.

5. Create an Audio Version of Your Newsletter

If you have a weekly newsletter, you’re already a content creator. However, inbox fatigue is real. An audio version of your newsletter, often called a private podcast, allows your subscribers to stay up to date without adding to their screen time.

For this to work best, it has to feel exclusive. It’s not a podcast. It’s a briefing. It should be under five minutes and focus on the “big idea” of the week.

6. Use Audio for Thought Leadership Distribution

Thought leadership is all about perspective. When you’re writing an opinion piece, you want to shift the reader’s mindset. And you can do that more quickly when they can hear the passion in your voice.

Take your most provocative or insightful LinkedIn posts and expand on them in short audio essays. Platforms like Swell are designed for these types of “micro-audio” conversations.

Source

Tip: Do these while the thought is fresh in your mind. Authenticity is more important than production quality in this format.

7. Build an Audio Resource Library

Think of it as an “Audible for Your Brand.” Eventually, you’ll build up your collection of narrated blog posts, interview excerpts, and essays. You can create a library section of your site.
It’s nice for a busy marketer to be able to “binge-listen” to something they’re interested in. If they’re a marketer who cares about SEO, it should be possible for them to click a button and start listening to all of your content related to search strategy.

8. Create Audio “How-To” Walkthroughs

If you have technical “how-to” guides or documentation, repurpose the steps into an audio walkthrough.

Instead of making a user tab back and forth between a help document and your software, give them an audio guide they can listen to while they perform the task.

This reduces friction and improves the user experience. It shows that you’ve thought about the practical reality of how people actually use your products.

9. Develop “Flash Briefing” Style Skills

For brands with a high volume of news-driven content, repurposing your “Top News” or “Daily Tip” section into a smart speaker skill, like Alexa Flash Briefings, is a powerful way to become part of a user’s morning routine.

Source

Once a listener subscribes, your 2-minute repurposed tip plays automatically alongside the morning news and weather.

It builds brand recall through daily, low-friction interaction.

10. Record Personable Q&A Responses

Static “Frequently Asked Questions” pages are usually where user engagement goes to die. They are functional but rarely memorable or persuasive. Audio responses turn a dry FAQ into a conversation.

Source

Take the top five questions your sales team hears every week and record short, personable audio responses.

Instead of just reading a paragraph, the user can hear the expertise and confidence in your team’s voice. This humanizes the problem-solving phase of the buyer’s journey and addresses objections with a personal touch.

11. Curate Audio “Roundups” of Industry News

Most industries move at a breakneck pace, and your audience likely struggles to keep up with the weekly firehose of updates. If you already write a “Weekly Wrap-up” or “Industry News” blog post, you have the perfect script for a high-value audio asset.

Instead of just linking to external articles, record a 5-minute summary that provides your brand’s unique take on why these news items matter. This shifts you from being a mere aggregator to a trusted curator.

Busy professionals can get their weekly industry briefing during their morning coffee or while checking emails. It positions your brand as the filter that helps them separate signal from noise.

Making Audio Work Without the Overhead

The biggest barrier to audio is the perceived effort. Marketers assume they need a soundproof room and an expensive microphone. In reality, the barrier to entry has never been lower.

The democratization of these tools means that small agencies and mid-sized businesses can now compete with global brands.

Whether you’re using Descript to edit audio as easily as a Word doc, or leveraging self-serve platforms like AudioGo to handle your podcast advertising, the technology is finally catching up to our creative needs.

Summary Checklist for Audio Repurposing:

• Audit: Identify your top 10 most-read blog posts from the last year.
• Narrate: Use AI or a voice actor to create audio versions of those posts.
• Clip: Break down long audio into 60-second “social snacks.”
• Promote: Use targeted audio ads to reach listeners on streaming platforms.
• Analyze: Track whether audio listeners stay on your site longer than traditional readers.

Your Turn to Speak Up

Audio isn’t replacing the written word; it’s amplifying it. By giving your content a voice, you meet your audience where they are—in their cars, at their desks, and on the move.

The real opportunity is in converting your best content into audio experiences that can be consumed on the go. This could mean narrated blog posts, audio insights, or audio ads to reach new audiences with your ideas.

Which of your current “hero” content pieces could translate well into audio?

Maybe you’ve heard…

The key to AI visibility is FAQs, schema markup and off-site citations on sites like Reddit. This is the consensus in the SEO community, standard advice on SEO articles and a common theme at SEO events.

To confirm my suspicion, I used a tool to scan 150 articles about AI search optimization on SEO blogs. Indeed, when categorized, these were the top recommendations.

I could easily have had the tool scan 500 or more articles, but the rankings stopped moving after the first 60. These top five categories were already locked in. Why have the AI burn the calories? More sources would have produced the same chart.

I worry that SEOs are often over-relying on checklists and tools without thinking creatively or strategically. It’s too easy to let best practices be a crutch for low performance. And in this case, I’m skeptical of the best practices themselves.

Before I skewer the typical SEO advice, let’s review the summary of the standard advice from SEOs on AI search optimization:

Typical SEO advice about AI search

1. FAQs & answer-first content (93% consensus in the dataset)
   Structure pages with Q&A headers, extractable answer blocks and dedicated FAQ sections so AI can easily parse and cite your content.
2. Schema & structured markup (89% consensus in the dataset)
   Tag your content with JSON-LD schema (Article, HowTo, Organization, Product) and configure llms.txt and robots.txt so AI crawlers can read it.
3. PR & third-party citations (87% consensus in the dataset)
   Earn mentions in press coverage, listicles, analyst reports and Wikipedia because 85% of AI brand citations come from third-party domains, not your own site.
4. Communities and UGC (79% consensus in the dataset)
   Build authentic presence on Reddit, YouTube, LinkedIn and review sites where AI engines verify brand claims against real community conversations.
5. Freshness & topical authority (77% consensus in the dataset)
   Update cornerstone content quarterly, publish original research and build topic clusters because AI citations decay after roughly 13 weeks.

Clearly, they’re all saying pretty much the same thing. I’m going to challenge that thinking now.

Here comes some unconventional wisdom about AI search optimization. I’m not saying my advice is original. I’m sure hundreds of SEOs do these things every day. But this is advice you don’t see in many articles.

SEOs suggest FAQs

…but don’t know which questions are frequently asked

Yes, AI users ask questions and AI gives answers. So it makes sense that the grammatical structure of questions and answers would help AI respond to users. The idea of adding FAQs makes sense.

The Problem: SEOs often have no idea what questions are frequently asked
How do SEOs know what questions to answer? The articles recommend SEO tools, Google, competitors and prompts.

They are mostly saying the same thing: just look up the questions, generate the FAQs, drop them into pages …move on to the next item on your checklist.

Look at the unusual advice at the bottom. Just look up the questions, generate the FAQs and drop them into pages. Easy, right? Move on to the next item in your checklist.

But you ended up with the same questions everyone else has.

The Fix: Ask AI to analyze your sales call transcripts
Use your own data to discover what questions your specific audience is asking. You probably have a lot of this first-party data. It’s your sales call recordings.

Post-covid, we’re all on virtual calls and many of these meetings are attended by AI notetakers. Indeed, I’ve been in meetings with more notetakers than humans! Jock and I had a lot of company that day.

Those sales call transcripts are huge piles of language directly from your target audience. It’s their problems and their questions in their words. A goldmine.

Export those transcripts. Put them all into a lightweight file (TXT or markdown) and upload them to your favorite AI. If they’re too long, break them up into multiple files. Better yet, make it a tool (custom GPT, Claude Project, Gemini Gem, etc.) so you can talk to it anytime. The insights are incredible.

I put 65 sales call transcripts into NotebookLM and it worked especially well because that tool sticks closely to the source material. It rarely hallucinates. It’s not as generative (AI engineers call this “low temperature”) which makes it bad for analysis but good for extraction. You get your prospects’ actual words back.

All loaded in? Use these prompts to find the real FAQ of your prospects.

• “What are the most common questions prospects ask in these calls? Categorize and give specific examples”
• “Analyze the prospects’ language to determine the most significant question they are thinking but hesitant to ask.”
• “What do prospects assume about us that’s wrong?”
• “What words and phrases do prospects use to describe their problem before they know the name for the solution?”
• “What specific criteria or requirements do prospects mention when comparing options?”

Voice of customer (VOC) is the gold standard in audience research and you’re sitting right on top of it. Call transcripts are a rich source of insights about questions and concerns. There is none better.

And it’s really all about your visitor in the end. Empathy is the entire point of a truly frequently asked question.

Answering the right questions isn’t just good for FAQs. It’s why you have a page and a visitor.

Yet, only three of the 150 articles analyzed suggested this approach. The most effective method is the least likely to be recommended by SEOs.

SEOs suggest schema markup

…even on pages with thin content

It’s easier for a machine to understand a page if the content elements are tagged. These tags are called “schema” and the schema.org website is a huge repository of all possible tags. Crawlers, including GoogleBot and GPTBot, can use these tags, so it makes sense to add schema markup.

The problem: SEOs recommend schema as a retrofit, not a content opportunity
The typical SEO advice is a workflow handled by technical SEOs and looks like this:

1. Audit your pages
2. Identify content that matches a schema type
3. Add the JSON-LD schema markup
4. Check it with a tool (Schema Markup Validator or Google’s Rich Results Test)

In other words, there is zero consideration for the content or the visitor. The page may have huge holes. Or it may have very little content at all. But the SEO articles all recommend the same approach: wrap what’s already there in tags and call it a day.

Your page can pass the schema test, but be useless at helping an AI understand your brand.

The solution: Use schema as a tool for planning content
Rather than thinking about schema as a tech SEO checklist item, handled after the page is live, use schema during the content creation process to give you new ideas for making your pages more complete.

Schema standards are a checklist of elements a complete page should contain.

For example, the “ProfessionalService” schema has properties for “serviceType,” “areaServed,” “hasCredential,” and “knowsAbout.” If your page doesn’t include those, you have a content gap.

Here’s a way to use the schema spec as a framework to discover what’s missing from your page. Copy and paste the URL into the end of this prompt and run it in your favorite LLM.

Schema-First Content Enhancer

You are a content strategist who uses Schema.org as a creative lens for improving web pages. Instead of auditing schema implementation, you use schema types and properties as a structured vocabulary for spotting content gaps. Use this process:

1. Identify the page’s primary topic and the most relevant Schema.org type(s). Retrieve the Schema.org page for each type before proceeding.

2. Check what AI engines currently surface for queries this page should answer. Note content that cited/competing pages include that this page lacks.

3. Review the properties for those schema types — especially ones the page doesn’t address. Think of each missing property as a potential content gap. Look at often-overlooked properties like step,  areaServed, review, offers, areaServed, knowsAbout, mentions, and mainEntity. Also consider related schema types that could apply if new content were added.

4. Deliver one output: Schema-to-Content Opportunities. For each gap identified, provide: A. The schema property that revealed it (with Schema.org link) B. The specific content to add — not “add a credentials section” but “add a paragraph covering [specific credential or certification relevant to this service] because competitors ranking for this query include it and visitors use it to evaluate trust” and C. where it belongs on the page and why a visitor benefits from it

Prioritize by impact. Lead with additions that most improve the page for a real visitor. Skip anything that wouldn’t change reader behavior or answer a question they’d actually have. Do not recommend markup or structured data changes. Assume implementation is already handled. Before recommending any schema type or property, retrieve its Schema.org page and confirm it exists — if you cannot retrieve it, do not include it.

Analyze this page: [Enter URL]

Now you’re looking at a list of holes in your page and suggestions for how to fill them. Set aside an hour and write some new pageblocks. You’re helping human visitors and AIs know more about you.

Which of these are the best things you can do today to help AI understand your brand?

1. Add little JSON tags to your code
2. Update your site so it explains precisely what you do, how you do it differently, whom you do it for and why it makes a difference

The answer is obvious. If there’s a way to add the schema to the new content, great. The content and the markup are born together. But even if you can’t easily add markup to everything, you still made a more helpful, more comprehensive page. That’s the fundamental idea behind AI search optimization.

Yes none of our 150 sources suggested this schema-first approach to content updates. The more effective method isn’t recommended in the SEO articles.

Lily Ray, one of the most cited voices in AI search frames it this way…

---

Lily Ray, Founder Algorythmic “A lot of AI search visibility advice reinforces itself. Unique, original research can naturally earn social media discussion on Reddit and LinkedIn, along with third-party mentions and citations as long as it’s compelling enough to spark conversation. That same content can be repurposed into YouTube videos, TikToks, and Instagram Reels. Structured data and clearly extractable content blocks should already be table stakes for SEO professionals.”

---

SEOs recommend Reddit, YouTube and other platforms

…but don’t know the buyers’ prompts or AI’s specific sources

Yes, off-site citations matter. AIs train on the entire internet, not your own website. When an AI response links to something, it’s called a citation. The most common citations are Reddit, YouTube and Wikipedia. So naturally, the SEO community tells everyone to get on Reddit, YouTube and Wikipedia.

The Problem: SEOs recommend starting with the platforms, not the buyer.
The conventional SEO advice is to review the list of top platforms AI uses and then go get active on those platforms. It’s a template strategy.

Yes, Reddit may be the most commonly cited source in AI responses. But that does not mean Reddit shapes answers for your category. Reddit (or Wikipedia or YouTube) really only matters if the AIs reference it when your specific buyer prompts. In other words, sources that drive AI recommendations are category-specific and prompt-specific.

The Fix: start with buyer prompts, not big platforms
I wrote a full methodology for this last week, including the specific prompts to use. Here’s a super short version of the process:

1. Generate synthetic prompts. Use AI to create a nice set of at least 10 realistic, commercial-intent prompts based on your actual ICP. They’re the questions your buyer types when they’re actively evaluating options.
2. Run those prompts in Google. Copy and paste them one at a time into Google’s AI (Gemini or AI Mode). See which brands surface and which specific sources the AI cites.
3. Archive the sources. Use a follow-up prompt to extract just the brands and source URLs from the responses.
4. Analyze and prioritize. Feed that archive into ChatGPT to identify which off-site source types recur (directories, review sites, trade publications, analyst reports)

It’s a multi-step, multi-prompt method that shows which specific sources are important for your buyer and your brand. Here are the instructions and all three prompts.

The final output practical mini-strategy for off-site AI visibility, tailored to your category, not a generic checklist. It will look something like this:

Once you know the most important sources for your specific buyer’s prompts, you can focus on the ones that actually show up in the analysis. If you’re a B2B brand, it’s probably not Reddit.

Only a few of the 150 articles I analyzed suggest anything close to this. Most just point to the popular platforms and say go. It’s mostly the same advice for every brand in every industry.

Let’s put it all together into a little infographic

You can see how many SEO articles recommend the same things and how few SEO articles recommend a more strategic, creative approach.

The difference is stark. And it’s a reminder that the goal is not to follow best practices. The goal is to drive results. Best practices, after all, are simply good hypotheses.

---

Bryson Meunier, Senior Manager of Holistic Search, Intuit “SEO is not one size fits all and anyone who implements SEO best practices without thinking about their business is doing it wrong. But remember that some of these best practices are specific recommendations from Google or ChatGPT. These are worth implementing.”

---

Where SEOs don’t agree… what is this work called?

There is one area where there is no consensus at all: the name of this type of marketing. I’ve been calling it AI search optimization, but it goes by many names, as we can see in our analysis of the 150 SEO articles.

It’s perfect irony. There is broad agreement in the SEO community about what to do, but zero agreement on what to call it. The keyword researchers must be losing their minds.

We hope that you try the ideas we’ve shared here.

Even more, we hope that these approaches encourage you to invent your own methods.

And finally, we hope that you share your own methods with the broader community, as we’ve done here.

We’re all in this together, learning and teaching as we discover what works in this fascinating new channel for marketing.

AI-powered tools are everywhere. They promise to analyze your funnel, build your content strategy, and write your marketing copy in a fraction of time.

But AI also gets things wrong. And the more complex the context, the higher the probability you’d fall for an answer that sounds credible but is actually false.

When it comes to website optimization, blindly trusting AI’s feedback, your gut feeling or whatever worked for someone else’s website can cost you real money (or your job).

But what if you could confidently tell which variant is stronger or which change has a higher probability to move the needle within minutes? No prompts, no AI double-checking required.

This post walks you through 7 neuroscience principles that directly affect conversion rates and shows you how to apply them to both copy and design so you can make smarter CRO decisions (reliable and fast, no tools involved).

But first, here’s what you need to know about the brain’s structure and information processing.

Brain science for marketers: What you need to know

The brain structures that matter for conversion

Ever heard someone mention the “prefrontal cortex” or “amygdala” and wondered what the hell they’re talking about?

Here’s a quick translation:

• Sensory cortex = processes sight, sound, touch, etc.
• Amygdala = emotional responses, threat detection, fear/safety signals. The immediate “something feels off here” alarm.
• Hippocampus = pattern recognition and memory formation.
• Language centers (Broca’s, Wernicke’s areas) = process words.
• Prefrontal cortex = executive functions including attention control, decision-making, working memory, and switching focus based on goals. Handles both rapid attention shifts and slower conscious deliberation.

Fun fact: In Greek, “Hippocampus” means “horse sea monster.” Hippocampi were the fish-tailed horses that pulled Poseidon’s chariot across the sea in Greek mythology. The Italian anatomist who discovered it ca. 450 thought it looked like a seahorse, so here we are.

How information processing works

Here’s the sequence your prospect’s brain goes through when they land on your page (whether they realize it or not):

1. Sensory input: They see your page, read your headline, notice the layout. The visual impression—clean or cluttered, professional or amateur, premium or cheap—forms instantly, before they’ve read a single word.
2. Pattern recognition: Their brain scans for familiar patterns and anything that looks off. Is your page prototypical of your niche? Is your pricing page structured in a similar way as dozens of pricing pages they’ve seen before?
3. Emotional evaluation: The instant “can I trust this company?” gut check. The split-second “something feels off” reaction and instant alarm.
4. Conscious processing: The prefrontal cortex starts doing the logical analysis. “Does this solve my problem? Is this worth the price?”
5. Decision: Reaching out or leaving. Often the conscious brain is justifying what the emotional brain already decided.

Now when someone says “emotional processing happens before logical,” you know that it actually means “the amygdala responds faster than the prefrontal cortex.”

One critical constraint that affects information processing

Your brain has two types of memory:

• Working memory holds information you’re actively using right now. It’s temporary and has limited capacity.
• Long-term memory stores information you’ve learned and can retrieve later. It has much larger capacity than the working memory.

When someone lands on your page, everything happens in working memory first. Every element—text, images, buttons, forms, navigation—competes for the limited space.

The total mental effort required to process information in working memory is called cognitive load.

Cognitive load fluctuates constantly based on how the brain interprets the information you present your prospects. Increased cognitive load can lead to decision paralysis, increased processing errors, or abandonment of the task entirely.

Ok, now you’re all set. Let’s talk about the juicy part of it: principles that affect cognitive load and how you can use them in your copy and design to increase your conversion rates.

Principle 1: Processing fluency

The neuroscience

When information flows smoothly, the brain interprets that ease as a safety signal. But when something requires effort to process, the amygdala triggers a subtle alarm—the emotional threat response—before the logical brain evaluates the offer.

This is called processing fluency (or cognitive fluency).

Why it matters for conversions

Research shows that easier-to-process information gets judged as more trustworthy, even when the actual content is identical. So, when your page is hard to process, by the time conscious evaluation begins, trust is already compromised.

Easy to process = feels right = trustworthy. Hard to process = feels off = risky.

How to increase cognitive fluency

In your copy:

• Use simple sentence structure. Shorter sentences with a clear subject-verb-object relation reduce cognitive effort. “Track your expenses” is easier to process than “Enable systematic tracking of all financial outflows across your organization”

Chunk information into digestible sections:

• Break long content into clear sections with headings, and keep paragraphs to 2-4 sentences. Each break allows working memory to process and consolidate information before moving on, rather than forcing continuous processing of one long stream.
• Use progressive disclosure. Show prospects only what they need at each step, revealing additional information as they move through the page. This goes for input forms, too. Multi-step forms that show 3-4 fields at a time create less cognitive load than one long form showing 15 fields simultaneously.

In your design:

• Group related elements visually. Place related items close together with shared styling (borders, background colors, or spacing). This leverages chunking—the brain processes the group as one unit instead of multiple individual elements.
• Do not center more than two lines of text. Centered text is harder to process because the brain needs to make an extra effort to find the beginning of the next line.
• Use clean layouts and clear visual hierarchy. Headings, subheadings, and distinct sections let prospects find what they need faster, unlike cluttered designs or dense, uniform text.
• Use high contrast between text and background. Any way that makes the text hard to read negatively impacts cognitive fluency.
• Add white space between sections. Visual breathing room separates distinct pieces of information, reducing the number of elements competing for simultaneous attention.

Two real-life examples

Example #1:

Codarity ran an experiment for one of their clients, where they tested a short, direct headline against one stacked with redundant descriptors.

The longer version made visitors work harder to process multiple similar phrases just to extract the main idea, increasing cognitive load. The shorter, clearer headline, which still retained the key message, resulted in 16.9% more conversions.

---

Dan Charles, Founder, Codarity “The pattern shows up in copy, in design, in structure. Simpler wins. Not because simple is lazy. Because the brain rewards messages it can process without effort. When a headline is clean and direct, it feels clearer. When it feels clearer, it feels truer. And when it feels true, it converts. The technically developed headline, packed with keywords and descriptors, makes the reader work. The simple one that speaks straight to the outcome they want doesn’t.That’s the test worth running. Strip it back. See what happens.”

---

Example #2:

When Expoze.io increased the contrast between text and background on their homepage, making it easier for their visitors to read and process the content, they saw a 40% increase in attention to key sections and a 25% lift in CTA clicks.

What seems like a trivial change addressed an important issue: text readability. So, it’s not surprising that making the text easier to read resulted in a substantial performance improvement.

(Source)

Key takeaway:

Easy to process = trustworthy. Hard to process = risky. Removing friction from comprehension should be your first priority.

Principle 2: Specificity

The neuroscience

The brain processes concrete language differently than abstract language.

Abstract language only activates language centers. You read the words, you understand their meaning, that’s it.

Concrete language (specific numbers, tangible outcomes, sensory details) activates sensory regions of the brain and creates mental imagery.

Brain imaging studies show that your brain treats imagined scenarios similarly to real ones: When you vividly imagine an experience, it activates many of the same brain regions as during a real experience, evoking an emotional response.

Why it matters for conversions

Emotions are the strongest drivers for action.

If your prospects can picture what you’re talking about—be it their current problem, your solution in action, or the outcome they’ll experience—they feel the emotional weight of that scenario as if it’s already happening.

“Better results” gives your prospects nothing—no imagery, no emotional response.

“5 new clients in the first week” makes the relief of hitting their quota and the satisfaction of early success feel tangible—before they’ve even considered signing up.

How to use pattern specificity to your advantage

In your copy:

Whether you’re describing your prospects’ problems, benefits of your offer, or how it worked out for your past clients, make sure your prospects can visualize what you’re saying:

• Use specific numbers. “Reduce meeting time by 40 minutes per week” beats “save time on meetings.”
• Be specific with context. “I write copy for B2B SaaS homepages and landing pages” beats “I write conversion copy.” Both are specific, but one paints a clear picture of what you actually do.
• Use names and specific details when talking about past work. “Slack reduced onboarding time from 3 days to 4 hours” beats “major companies saw faster onboarding.”

In your design:

• Show actual product screenshots, not abstract illustrations. Prospects need to see themselves using it.
• Use before/after photos with real results. Visual proof of concrete outcomes reinforces what the copy promises.
• Display charts with actual data points, not decorative graphics. If you’re claiming results, show the specific numbers visually so prospects can picture the positive development.

A real-life example

FreshBooks noticed visitors were exploring their Product Tour and Features pages but not signing up. They tested a clearer, more specific version that:

• Highlighted the 5 features users cared about most (instead of listing everything)
• Added concrete context: “get you paid 2x faster” instead of vague “get paid faster”
• Explained what features actually do: “Automate invoicing and watch FreshBooks instantly send them to clients on your behalf”
• Made the product screenshot larger so visitors could see the actual interface
• Added a specific “Try it free” CTA instead of vague “Learn more about invoicing”

Once people could picture the specific outcome instead of processing abstract promises, sign-ups increased by 4%.

Control:

Variation:

(Source)

Key takeaway:

If they can’t imagine it, they won’t buy it. Replace vague descriptions with specific situations and outcomes your prospects can visualize.

Principle 3: Pattern recognition & expectation

The neuroscience

Pattern recognition happens automatically in the hippocampus and sensory cortex. Matching stored patterns requires minimal cognitive effort. Violating expected patterns forces the brain to slow down and consciously analyze what’s different.

Why it matters for conversions

Unfamiliar patterns increase cognitive load. Your prospect’s brain has to work harder to understand what they’re looking at, which pulls cognitive resources away from evaluating your offer.

Breaking patterns strategically can work in your favor, for example, when using an unexpected color or size for your primary CTA makes it stand out and captures attention. But breaking patterns in infrastructure that should support information flow (navigation, forms, standard UI elements) just creates unnecessary friction.

How to use pattern recognition & expectations to your advantage

In your copy:

• Match expected category language in your header section. If someone searches “project management software,” make it clear that’s what you offer in the headline, subheading, or tagline. Don’t make them decode “Transform collaboration paradigms” to figure out you sell project management software.
• Stick to conventional navigation labels.  “About,” “Services,” “Pricing,” “Contact” require zero mental effort. “Who We Are,” “How We Help,” “Let’s Connect” make visitors pause and decode.
• Use the same words your audience uses. “I wish I could achieve customer acquisition enhancement”, said no prospect ever. But they do want to get more clients.

In your design:

Unless you’re strategically breaking a pattern to make something stand out (like an oversized CTA), stick to prototypical elements—standard UI conventions that match what people have learned from every other website they’ve used.

• Use prototypical navigation placement. Logo top-left. Navigation across the top or left side. Search icon in the top-right corner.
• Follow standard color conventions. Blue for links. Red for errors or urgent actions. Green for success messages.
• Make clickable elements look clickable. Buttons need borders, background color, or hover states. Input fields need visible boundaries and placeholder text. Text should look like text, not like a button.

Keep in mind expectations specific to your context

What elements show up on nearly every site in your niche? What sections do they all have? What functionalities do they offer? That’s what your prospects expect from websites like yours. Make sure to fulfill them as well.

A real-life example

Teamwork’s comparison page lacked the one thing visitors expected to find: a side-by-side feature comparison table. Instead, people had to jump back and forth between sections to compare Wrike and Teamwork feature by feature.

When GetUplift redesigned the page, they added a standard comparison table that matched what people had learned to expect from seeing dozens of comparison pages across the web. As a result, conversions increased by 54%.

---

Talia Wolf, Founder and CEO, GetUplift “After running thousands of A/B tests and observing how people behave on real webpages, we’ve identified 4 types of readers on every page. Each has learned specific patterns for how to consume content: where to look for proof, what headlines should do, how subheads should work, etc. When pages violate these learned patterns, conversions drop simply because people can’t find what they expect to find where they expect to find it.”

---

Key takeaway:

Identify what needs to be familiar (navigation, page layouts, form structure) versus what should be distinctive (your offer, key benefits, primary CTA). Make infrastructure invisible. Make your message stand out.

Principle 4: Attention & the Von Restorff effect

The neuroscience

Your brain is wired to notice what’s different. When you see a list of similar items with one that stands out, that distinctive item gets more attention and is remembered better.

The prefrontal cortex automatically detects the difference in context, triggering enhanced memory encoding that makes the distinctive item easier to recall later.

This is called the Von Restorff effect or isolation effect.

Why it matters for conversions

On your webpage, this means one distinctly different element will dominate where your prospect’s eyes go and what they remember.

Think strategically: what do you want to stand out? Your primary CTA? A key benefit? That’s what should be distinctly different. Everything else should blend into a consistent background.

How to use the Von Restorff effect to your advantage

In your copy:

• Use numbers in headlines when surrounded by text. “Reduce costs by 47%” stands out in a text-heavy section because the number breaks the pattern.
• Use power words strategically. If your copy is straightforward and factual, one emotionally charged word (“devastating,” “secret,” “forbidden”) will pop.
• Vary paragraph length. One short paragraph surrounded by longer ones stands out and signals emphasis.
• Use visual hierarchy to guide attention. Size, color, and position should make it obvious what to look at first, second, third, making the decision as to what to read next intuitively easy.

In your design:

• Create color contrast for your primary action. If your page is blue and white, your main CTA should be orange or green—distinctly different from the base palette.
• Use a single icon or graphic among text-only sections. When surrounded by paragraphs, one relevant visual element draws the eye and breaks the monotony.
• Add a background color to one section. If your page sections flow on white background, put your key offer or CTA section in a light colored box to make it visually stand out from the continuous scroll.
• Reduce animation and movement. Moving elements, auto-playing videos, or animated banners compete for attention simultaneously with static content, increasing cognitive load. Use motion sparingly and purposefully.

A real-life example

AliveCor added a “New” badge to their product, KardiaMobile Card, on both the product listing and detail pages. The badge created visual distinction, making one element stand out while everything else remained consistent.

Result: 25% increase in conversion rate and 30% increase in revenue per user.

Control:

Variation:

(Source)

Key takeaway:

When everything stands out, nothing stands out. Choose one element per screen to dominate attention—make only that one distinctly different. Keep everything else visually consistent.

Principle 5: Loss aversion & the pain-pleasure axis

The neuroscience

Our brain feels losses approximately twice as intensely as equivalent gains. This isn’t a thinking preference, but a hardwiring.

Originally discovered through behavioural studies, the neuroscience studies that followed showed that losses and gains are processed by different neural circuits, with losses triggering stronger and more widespread activation than equivalent gains.

Why? The scientists’ best guess: our brain evolved in a world where losing your food stash or your mammoth-hunting gear meant death, but finding extra berries just meant… extra berries.

Avoiding loss has mattered much more than chasing gain since long before we invented the internet, and your brain hasn’t updated its operating system.

Why it matters for conversions

Loss-framed messaging activates the amygdala before the logical brain evaluates your offer. “Stop losing 20 hours per week to manual reporting” hits harder than “Save 20 hours per week.”

The pain of continuing their current situation (so called “status quo cost”) motivates action more effectively than the promise of improvement. Your prospects are already experiencing the loss. Your job is to make them feel it.

How to use loss aversion to your advantage

The basic idea is either to make your prospects aware of an actual loss or to remove the fear of a potential loss.

In your copy:

• Frame benefits as loss prevention. “Stop losing customers to slow response times” works better than “Improve customer retention.” Both promise the same outcome, but one emphasizes avoiding loss.
• Quantify the status quo cost. “Your current process costs you $4,800 monthly in wasted hours” makes the pain concrete and urgent.
• Use scarcity and urgency strategically. “3 spots left” or “Offer ends Friday” creates fear of losing the opportunity. But use it only when it’s genuine. False scarcity destroys trust faster than it creates urgency.
• Include risk reversal. Money-back guarantees remove the fear of potential loss.

In your design:

• Use red or orange for urgency elements. These colors signal alert and danger, reinforcing the emotional weight of potential loss (unless orange or red are your primary theme colors).
• Place social proof near loss-framed copy. When you’ve named a pain point, immediately show testimonials from people who escaped that same pain. This proves the loss is avoidable.
• Use progress indicators in multi-step processes. “Don’t lose your progress” is a powerful motivator to complete forms or checkouts. If you ever had to pretend to learn a language for months just because a green owl threatens you with a loss of your 253-day streak, you know how powerful this is.

A real-life example

Leadforce tested a popup for Babuwear that triggered two loss-aversion signals: “stock may run low soon” and “here’s how much you’re saving.” The popup made the potential losses visible, creating urgency without relying on fake scarcity.

Result: 24.5% increase in conversion rate.

(Source)

Key takeaway:

The brain responds more powerfully to avoiding loss than achieving improvement. Name clearly what your prospects are currently losing, not just what they could gain, and frame your value as loss prevention.

Principle 6: Anchoring

The neuroscience

The first piece of information you encounter, the anchor, becomes the reference point for evaluating everything that follows. The prefrontal cortex uses this initial anchor to make rapid comparisons and value judgments.

Why it matters for conversions

Your prospects don’t evaluate your offer in a vacuum. They anchor to the first value signal they see, whether that’s a competitor’s price they saw yesterday, the “original” price you crossed out, or the first benefit you mention.

Present a high anchor first, and your actual price feels reasonable. Start with a low-value anchor, and even a good price seems expensive.

Control the anchor and you control how everything else gets judged.

How to use anchoring to your advantage

In your copy:

There are at least 6 types of anchoring you can use:

• Price anchors: “Most agencies charge $15,000 for this” sets the anchor. Your $6,000 price now looks like a deal.
• Time anchors: “This usually takes 40 hours to learn” makes your “master it in 3 hours” course feel ridiculously fast.
• Performance/results anchors: “Industry average response time: 48 hours” makes your “4-hour response” feel exceptional.
• Effort anchors: “With the common solution X, you need dedicated IT staff” makes your “no technical knowledge required” feel like a jackpot.
• Social proof anchors: “Trusted by Google, Microsoft, Amazon” anchors quality and legitimacy before prospects even know what you do.
• Feature/benefit anchors: The first benefit you mention becomes the baseline for everything else.”50GB storage, 10 integrations, saves 5 hours/week” anchors on “50GB” as the baseline value. When you get to “saves 5 hours/week,” the number feels small, and just like another item in the list, although it’s as valuable as (if not more than) the rest.”Saves 5 hours/week, plus 10 integrations, plus 50GB storage” anchors on “saves 5 hours/week” as the main value. Everything else feels like bonus features on top of that core benefit.

In your design:

• Place the highest price visibly first in price tables. Even if it’s not your recommended option, it sets the anchor. Everything else looks more reasonable by comparison.
• Show the “original” or “list” price crossed out next to your actual price. The crossed-out $199 anchors the value. Your $99 sale price now feels like you’re getting $199 worth of value for $99.
• Display competitor prices or industry averages in comparison charts. Visual anchors work powerfully. When prospects see “$2,500 (competitors)” next to your “$899,” the $2,500 anchors their perception of value.

Why do most SaaS pricing pages list packages low to high?

To anchor or not to anchor depends on your business model.

A typical business model of SaaS—get them through the door with the cheapest plan first, upsell later—outweighs the anchoring benefit of showing the expensive option first.

A real-life example

When testing a landing page for his ebook, Michael Aagaard from Unbounce used anchoring to test two different variations of a value-packed bullet point.

The original sentence anchored on credentials: “Insights and experience from 4 years of research and over 350 A/B tests distilled into one 26-page free ebook.”

The variation flipped the order to lead with accessibility: “Read the book in just 25 minutes and get insights from 4 years of research and over 350 A/B tests.”

By leading with the low time investment (25 minutes), the ebook felt like a quick read, whereas, starting with comprehensive credentials made readers anchor on substance and depth, signaling a bigger time investment.

This simple change resulted in an 18.6% increase in downloads.

(Source)

Key takeaway:

The first number or value claim your prospects see becomes the anchor for evaluating everything else. Decide on what to show first strategically to make your offer look like the best choice by comparison.

Principle 7: Social proof & conformity bias

The neuroscience

Uncertainty activates brain regions associated with conflict detection and anxiety, which can trigger avoidance behavior or decision paralysis.

When faced with uncertain decisions, the brain looks to what others have done. This conformity bias is most probably a tribal survival mechanism: if five of your tribe members ate these unknown berries and survived, you can probably eat them, too.

Why it matters for conversions

Every purchase decision involves uncertainty. Will this work? Is it worth the money? Can I trust this company?

When your prospects see that others—especially people similar to them—made the same choice and got results, this lowers their brain’s uncertainty signal. The risk feels lower, and the decision becomes easier.

A note of caution: not every testimonial helps reduce uncertainty

Generic praise (as in “We highly recommend this company”) won’t make your prospects trust you more.

An effective testimonial needs specific, relatable details—job titles, company names, industry context—so your prospects’ brain has enough information to assess whether the results it describes would apply to their situation.

Video testimonials are most powerful because they activate face recognition, facial expression reading, and vocal tone processing—systems your brain trusts to reliably detect authenticity.

How to use social proof strategically

In your copy:

• Include testimonials from people similar to your prospects. The more they can see themselves in the testimonial, the stronger the effect.
• Display usage statistics that demonstrate adoption. “Join 50,000+ marketers” signals that many people have already made this choice. The larger the number, the lower the perceived risk.
• Feature authority endorsements when relevant. Quotes from recognized experts or publications add credibility, especially if your prospects need validation beyond peer testimonials.

In your design:

• Make client logos immediately visible. Recognition triggers trust before prospects read a word. Place logos prominently, not hidden at the bottom of the page.
• Display star ratings and review counts. Keep in mind that volume and authenticity matter more than perfection. “4.8 stars from 2,400 reviews” works better than a perfect 5.0 from 12 reviews.
• Show numbers visually. Graphs showing customer growth, charts displaying satisfaction ratings—visual proof reinforces what your copy claims.

A real-life example

Vegetology had customer testimonials on their product pages, but they were buried at the bottom where visitors rarely scrolled. They tested moving a testimonial above the fold, visible immediately when someone landed on the page.

The change put social proof exactly where uncertainty was highest: the moment someone was evaluating whether to trust an unfamiliar store. Result: 6% increase in conversions.

(Source)

Key takeaway:

The brain treats “people like me succeeded” as proof of safety. Show relatable people who made the same choice and got results to reduce uncertainty in decision-making.

When multiple principles contradict each other

But what to do if making things simpler reduces trust? Or when adding more information results in lower processing flow?

Here’s when knowing your target audience becomes indispensable.

If you know what matters to them most while making a buying decision you’ll know what principle to choose when deciding on the most promising variants to test.

For example, in this A/B test on their website, TruckersReport, a network of professional truck drivers, learned that reducing the number of input fields in their form leads to reduced conversion rate. The control version with 4 input fields outperformed the variation with just one field by 13.56%.

It seems like their target audience valued additional relevancy and credibility (“if they’re asking for my location and driving experience, their offers will be more suited for me”) more than having just one field to fill.

(Source)

All principles at one glance

Here’s a quick reference of all principles and mechanisms behind them to make it easier for you to put them into practice.

How to apply these principles to increase your conversion rates faster

Pick one high-traffic page. Read through it while asking:

1. Is this easy to process? (Fluency + load)
2. Is this specific and clear? (Specificity)
3. Does this feel familiar? (Patterns)
4. Does the important stuff stand out? (Attention)
5. Do I address their pain? (Loss aversion)
6. Do I prove I’m credible? (Social proof)

Identify your 2-3 biggest violations. Then create test variants that fix them.

You’re now optimizing a strong foundation.

Don’t have enough traffic to test? Now you can make an educated guess as to which change will have the biggest effect.

Beyond conversion rate optimization

Understanding these principles doesn’t just optimize your web pages. It changes how you approach every email, every presentation, every sales conversation.

Because you stop guessing what might resonate or defaulting to what “sounds good,” and start building on what your counterpart’s brain is wired to respond to.

Your future prospect is doing a little research, looking for a new partner.

Who do they find?
Do they find you?
Did AI recommend your brand?

The answer is partly shaped by sources that aren’t your website. Yes, your website is your first and best opportunity to train the AI to recommend your brand, but AI trains on the entire web. So off-site sources matter.

Before you go off-site, make sure your site is ready. Use Orbit’s free AI Readiness Assessment to score your website across the key dimensions that shape AI recommendations then come back here for the off-site strategy.

SEOs all know this and we’ve all seen the research that AIs cite some websites more than others. Maybe you’ve heard, Reddit is one of the most commonly cited websites in AI responses. Here’s some data from Profound.

But does that mean every brand needs a Reddit strategy? I’m not so sure.

Just because Reddit (or YouTube or Wikipedia) is a common citation doesn’t mean the AI relies on it to respond to your future prospect when they prompt.

Reddit (or any source) only matters if it AI checks it when your buyer searches for brands in your category.

The sources that influence AI responses are category-specific and query-specific. 

Saying every brand needs a Reddit strategy is like saying every brand needs a Facebook strategy. Facebook is popular therefore every brand needs to be on Facebook. Of course, that’s absurd.

Don’t chase every visible citation source. Instead, identify which external sources consistently shape AI answers for your buyer’s actual use cases, then prioritize the channels that you can realistically influence.

So don’t start with the sources. Start with the prompts.

Today, we’ll peek into the AI training data to see which sources appear to shape AI answers. You’ll learn that those sources are discoverable and practical to assess.

Do to that, we’ll use a four-step process:

1. Predict the prompts that your buyer uses when researching your category (ChatGPT)
2. Run those prompts in Google (Gemini or AI Mode)
3. Archive the responses and sources for those responses (Gemini or AI Mode)
4. Analyze the sources that influence responses and prioritize next steps (ChatGPT)

Why can’t we just do this all in one model?
You can. It can all be done in Gemini but the analysis isn’t as good. Better to use two models, where one analyzes the other.

Doesn’t our prompt tracking tool do this?
Maybe. It has the reports, but not the analysis. We set up prompt tracking for all of our optimization clients. If you have one of these tools, export the top source URLs (exclude your competitors) and skip to prompt four below.

When you’re done, you’ll have a mini-prioritized strategy for off-site AI search optimization. There’s a screenshot below of what it will look like.

Step 1: Generate Buyer-Specific Commercial-Intent Prompts

This prompt predicts the likely “buyer-intent prompts” which we need to uncover which off-site sources shape AI answers. It creates realistic buyer-intent prompts that sound like something a real prospect would actually type into an AI tool while evaluating brands like yours.

The inputs are very important. You need to tell it about your buyer. If you give it the right details, the predicted prompts will be more accurate. If you have defined ICPs, personas, keyword data or market research, pull from those.

• Buyer profile: Job titles, company size, geography and anything else
• Buyer industries: The verticals they work in
• Additional context: What problem are they hoping to solve? What sent them looking for help? What do they already know about what they need?

Pro Tip: If you’re struggling to fill this out, try this prompt to get you started: “Visit [website] and infer the most likely ICP. Then list the buyer profile, industry and additional context. Keep the total response under 90 words, use compact phrases (no paragraphs) and skip the explanation and commentary”

Review carefully, edit, then plug those details (probably 100 words at least) into the top of this prompt.

Buyer-Specific Prompt Generator Prompt (ChatGPT)

Generate 10 buyer-style prompts most useful for identifying which off-site sources appear to shape AI answers in this category.

Assume the buyer already understands the category and is actively evaluating vendors. Exclude educational, exploratory, and trend-based prompts.

Inputs:

• Buyer profile [job titles, company size, geography]
• Category: [their industry] 
• Additional context [what triggered the research, primary concern, what they already know, what they don’t know]

Choose prompts that span key buying stages such as discovery, shortlist, comparison, validation, implementation risk, and ROI.

Write each prompt like a real buyer question: short, natural, and commercially specific, usually under 12–15 words. Avoid long setup, extra detail, and multi-part instructions. Include only details a buyer would realistically mention, and prioritize variety over near-duplicate comparisons.

For each prompt, include:

• Copy/Paste This Prompt
• Buyer Intent
• Decision Criteria Reflected
• Likely Source Types Surfaced

Append this instruction to every “Copy/Paste This Prompt” entry:

Use current web information. After answering, include:

• Brands mentioned
• Third-party sources referenced
• A plain list of specific supporting links or publications, one per line

Also include a short explanation of why these prompts were chosen, brief notes on likely source diversity across the set, any missing or overrepresented stages, and any prompts intentionally deprioritized.

Formatting rules: The response should italicize the text in the Copy/Paste This Prompt section and put it in quotes. Copy/Paste This Prompt must appear first for every prompt and its content must be a single clean paragraph, not bullets. The user should copy and paste only the text in that section into another AI tool. Keep the other fields concise and easy to scan, and keep the output clean and easy to review one prompt at a time.

The prompt should look something like this. Those details are used to predict the prompts your buyer uses to research your category.

The response includes 10 prompts that your buyer would likely use to find companies like yours. Within each prompt is a small additional instruction that will make the sources easy to extract for off-site AEO analysis.

It will look something like this:

Notice that the prompts don’t include any brands. We want to see how it responds, who it finds and what sources it uses when we start fresh. We are emulating the experience of a non-brand aware buyer.

Now you’re ready to proceed with the 10 “prompt runs” in Google.

Step 2: Ten “prompt runs” in Gemini or Google’s AI Mode

We recommend Google because it is dominant and a popular B2B research tool for buyers. Gemini may not be the most popular AI chatbot, but it’s catching up fast. Google Search is pushing AI hard (AI mode and AI overviews) so it’s likely to become even more important. But you could use this method for any LLM. Try them all, if you’d like.

Now comes the tedious part.

Copy and paste the first of the 10 prompts (the highlighted area in the screenshot above) into Gemini or Google’s AI Mode. Send it. Then copy and paste in the second. It will be in one long conversation. Each prompt and response will look something like this:

Certainly, there are more comprehensive ways to do this audit and you can absolutely take this method further if you have time.

• Use a larger sample: Do more than 10 prompt runs
• Use more LLMs: Go beyond Google and run prompts in other models
• Do prompt runs in multiple, separate conversations: Doing it all in one conversation may create an “AI momentum” problem where subsequent responses are affected by previous prompts, causing the results narrow and possibly missing important sources.

But we don’t need a comprehensive audit to spot the patterns and get quick ideas for action. This is enough to proceed with our analysis.

---

Britney Muller, Founder, Organe Labs “The ’10/10 runs’ approach is a solid instinct, because AI outputs as you know are non-deterministic. The same prompt can give you different answers each time. Ten runs give you a better, but still a very crude directional signal. It’s really not statistical certainty.”

---

Step 3: Create an archive of the responses and sources (one final prompt in Gemini)

Now that you’ve completed the 10 “prompt runs” in Gemini or AI Mode, it’s time to pull out the sources. We’ll use another prompt to distill down just the details we want: the prompts, the brands in the responses and the sources the AI used.

This prompt will put them all together into a simple plain text archive that you can copy to your clipboard. Just enter the prompt below into the same conversation with Google, right after the last of the 10 prompt runs.

Note: This works perfectly in Gemini, but in tests within Google AI Mode, I had to run this prompt twice. That tool is really an extension of a search engine, not a typical generative AI. If it doesn’t work the first time, just try it again.

Source Archive Generator Prompt (Gemini or AI Mode)

Extract the off-site source evidence from each previous buyer-intent response in this session.

Start the output with this exact title on the first line: “Off-Site Source Evidence Archive”

For each response, create a clear “RUN [Number]” section and include only:

• Prompt
• Brands mentioned
• Third-party sources referenced
• Third-party links

Do not summarize, rewrite, or add analysis. Copy the content as written where possible. Preserve all links. Return the final output in a single code block.

This prompt removes all the noise and provides you with just what you need: the prompt, the brands that were surfaced and the sources it used to surface those brands. Now it’s all in a neat little archive. Perfect for our analysis.

Copy and paste the archive to your clipboard.

Step 4: Discover which off-site signals show up the most (and what to do next)

Our final prompt takes the archive and identifies, categorizes and prioritizes the recurring third-party sources. These are the off-site AEO signals that inform Google’s recommendations when your buyer researches your industry.

It also puts the audit in the context of your own brand. Did you show up in the responses? To find out, enter your brand at the top of the prompt.

You could run the final prompt right in the same conversation with the prompt runs (Google Gemini or AI mode), but in our tests, the analysis was better in ChatGPT. So we recommend going back to ChatGPT and continuing the conversation you started with the first prompt above. I’m sure Claude would also be great for this.

Here’s the prompt. It’s a doozy. Paste in the source archive from your clipboard along with this prompt.

Off-Site Source Influence Audit (ChatGPT)

I’m pasting an archive of completed buyer-intent AI responses below. Analyze it to identify which off-site sources appear most likely to shape AI answers in this category.

My Brand: [company name]

Treat this as an observed-pattern audit, not a definitive map of model influence. Base conclusions on recurring surfaced sources, recurring source types, repeated brand visibility, and repeated patterns in how evidence is presented across the runs. Use My Brand as a secondary interpretation lens, not the primary basis for the audit.

The archive is the primary evidence base. It may be inconsistent, repetitive, or uneven in evidence quality. Some cited items may be standards, agencies, regulations, analyst firms, review sites, publications, directories, or vendor pages. Some “third-party sources referenced” may be inferred in the response rather than clearly evidenced by linked support, so distinguish observed evidence from inference.

Return:

1. Key patterns

Briefly summarize the most important recurring source, source-type, and brand patterns across the runs.

2. Off-site source priority table

Create one markdown table ranking the top 5 off-site source categories most likely shaping AI answers in this category.

Use these columns:

– Ranked Category

– Example Sources

– Why It Matters / Confidence

– Recommended Off-Site Actions

In Ranked Category, combine the rank, category, and visibility into one cell using this format:

[number]. [source category] ([x]/[y] runs)

In Example Sources, list 3–6 specific named sources from the archive whenever possible, such as actual directories, publications, review platforms, analyst firms, associations, or partner ecosystems. Prefer exact source names over generalized labels. Do not use vague labels like “category-specific directories,” “verified review platforms,” “news-oriented ranking posts,” or “platform ecosystem references” if a specific named source is available in the archive. Only use a generic source label if the archive clearly suggests a source type but does not name a specific source. Format the contents as a short bullet list within each cell so the examples are easy to scan.

In Why It Matters / Confidence, start with a short confidence label such as High, Medium, or Low, then briefly explain why the category matters. Rank categories by likely off-site influence and recurring visibility across runs, not by confidence alone.

In Recommended Off-Site Actions, give only practical off-site actions that increase third-party visibility and credibility in AI answers. Prioritize analyst outreach, directory inclusion, review generation, trade PR, association participation, awards, partner amplification, and community presence. When possible, make the recommendation source-specific so the next step is obvious. Reference the named sources in that row rather than giving only generic advice.

3. Competitive readout

Briefly summarize which brands appear most often, which seem most supported by third-party signals, any smaller brands that appear to overperform, and any brands that seem boosted mainly by brand-owned citations rather than third-party support.

4. Brand gap readout

Using My Brand as the reference point, briefly summarize how often My Brand appears across the runs, which off-site source categories seem to support My Brand most, where My Brand appears underrepresented versus competitors, and the top off-site opportunities most likely to improve My Brand’s visibility.

5. Evidence quality notes

Briefly note patterns that may weaken confidence, such as repeated vendor-owned links, repeated uncited claims, standards used as credibility signals without direct evidence, low-quality sources, or duplicated source patterns.

6. Prioritized action plan

Give a short prioritized action plan based on the observed off-site source patterns and My Brand’s visibility gaps. Include the top 3 highest-impact off-site actions to take first, why each action ranks where it does, the expected visibility benefit of each, and any important dependencies or sequencing.

Use the archive silently as background input and do not surface the pasted file, file name, badge, attachment, or source chips anywhere in the output. Focus recommendations on off-site actions only, not on-site publishing, website updates, or brand-owned content as primary actions. Prioritize recurring third-party signals over one-off mentions, and treat brand-owned and competitor-owned sites as contextual rather than primary evidence.

Here’s what the analysis will look like:

These are the sources that influence AI recommendations in your category.

Suddenly, this key component of your AI search strategy becomes more clear. This mini-report lays out next steps and which internal team or outside partner may do the work.

• If AI is using review sites to train on your category…
  you can influence responses through outreach to fans and reputation management. (usually handled internally as part of customer marketing)
• If AI is using directories to train on your category…
  you can influence responses through submission and management (usually handled by the in-house marketing team, so you’ll have the logins and can manage your listings forever)
• If AI is using trade pubs and associations to train on your category…
  you can influence responses through content marketing collaborations and influencer marketing (usually handled in-house, but can be outsourced to a PR firm)
• If AI is training on analyst reports to train on your category…
  you can influence responses through outreach and requests for inclusion (usually handled in-house, but can be outsourced to a PR firm)
• If AI is training on “listicles”to train on your category…
  you can influence responses through outreach or guest posting (often handled SEO or digital PR firm)

Notice how some of the actions could be best handled by a partner, such as an SEO agency or PR firm, and some are probably best done in-house.

I suspect you get a lot done in-house before outsourcing. If you’d like an action plan with detailed playbooks for each, just ask the AI in the next prompt. Tell it all about your internal resources, current partners and budget.

---

Corey Northcutt, Chief Optimization Officer at Orbit “Most of the popular GEO myths — the list of which already seems endless — are about “off-page” work. There are no “signals” like with SEO. No scores, no math, no webspam counterweights. Not yet, anyway. Where AI results are not driven by organic search, it’s purely about word frequency and proximity. Here, the most biased result almost always wins. Look at a few hundred of your own prompts, and you’ll quickly find Reddit and Wikipedia citations to be less noticeable than in Google. Instead, it’s a small, actionable list of niche 3rd party affiliate websites, sponsored endorsements, or other biased actors that you can easily prioritize.”

---

It’s true that AI responses aren’t always based on searches. That brings us to our final topic today…

What does AI know about your brand from memory without searching?

We just made the case for very targeted efforts (directories, review sites, listicles, analysts) but what about all of the other sources? What about news media and mainstream press? What about press releases? What about Forbes? What about those high-domain authority websites that SEOs love?

There is definitely a place for traditional PR in your AI-search strategy.

Because sometimes, AI doesn’t search. Sometimes it just relies on its pre-training, spots certain players and recommends them. Big brands are big in AI responses because they’re big in the pre-training. Remember, the “P” in ChatGPT stands for “pretrained.”

What does ChatGPT know about your brand from its pre-training?

There’s an easy way to find out.

You could simply ask ChatGPT about your brand and tell it not to search, but that isn’t reliable. It’s hard to stop it from searching. The best way to look into ChatGPT’s memory is to make a custom GPT with “Web Search” turned off.

We made one that you can use Orbit’s No-Search Brand Visibility GPT.

It’s very simple. Just enter your brand and find out what AI remembers about it without searching. It’s a strange tool because it’s LESS functional than the default AI. We removed one AI’s main features!

This is not a perfect look inside pre-training, but it reveals what the model says when live web search is not available. It’s a clean, non-search test. It must rely on what it already knows.

And if it doesn’t know much, traditional PR will help. Press placements are huge in the training data. Likely, reputable media is more heavily weighted than company websites. And the way to become famous in the mind of ChatGPT is the same as it is offline: tell your best stories through credible sources. That’s PR.

Colophon

The prompts in this article were created using our own “Prompt Collaborator GPT” over the course of eight hours and dozens of exchanged messages. Rather than asking the AI to simply generate these prompts, I started with a long conversation about the general concept. I asked many open-ended questions. Many times, the AI provided valuable ideas for improvement. This final version was consolidated from five prompts down to three, then we tweaked and tweaked so the outputs are practical and nicely formatted. Tests were run across multiple brands, industries, and AI models.

This method is almost like a little piece of software. Indeed, it may partly replace tools you’re paying good money for.

Someday, we’ll make an article and video showing how to build your own prompt collaborator and work with it to create your own highly functional, multiprompt AI methods.

For some insights, you just have to ask people. Surveys are only way to answer questions like these:

• When do people prefer traditional search over AI chatbots?
• Do people find Google’s AI Overviews useful?
• What do you trust more? Google or AI chatbots?

To answer these questions and others, we partnered with our friends at the survey software company, QuestionPro. Together we surveyed 1,110 people in the US from all 50 states, asking about their preferences and use of AI chatbots and search engines.

Today, here in March 2026, we are sharing the results, and showing the changes since last year. We’ll answer six questions with six charts.

1. Are people switching from search to AI chat tools?

Tech changes fast. But what about user behavior? Are users ditching the traditional search experience and going straight to an AI chatbot?

Some users have already switched completely. Others, not at all. You may know people who have replaced Google with ChatGPT or Perplexity. Today, more than half of respondents start a search by opening an AI app.

But that number isn’t rising. Use of Google hasn’t declined. It’s no surprise considering the popularity of Chrome (51% of US internet users) and Google as the default search engine for Chrome. And of course, Google is the default search engine for both Android phones (through Chrome) and iPhones (through Safari). Every US phone pushes users to Google Search when they need something. Using an AI chatbot is harder. It requires that the user installs an app.

Here’s how Claude summarized this data: “AI-first enthusiasm is moderating into more selective use.”

2. When do people prefer AI for searching?

If users are being more selective, then when do we use which tool? Search engines or AI chatbots?

It depends on the intent of the searcher. Every SEO thinks about intent, and knowing which keyphrases have which kind of intent is the key to winning at search. Mostly, search pros think about two types of intent:

• Information intent: They just want an answer. They’d prefer not to visit a website. Google mostly provides the answer in the search results, so clickthrough rates are mostly down for these phrases. If they do click, conversion rates are low.
• Commercial intent: They’re researching a decision. They’re looking for a product or service. Clickthrough rates are higher for these phrases and when they do click, conversion rates are higher.

For the survey, we broke up the options into more specific types of intent with sample queries. The use of Google vs. AI chatbots varies by query type. And it’s gradually changing:

AI is becoming more popular for all types of queries, even the simplest ones. But the exception is local search, which makes sense because AIs don’t integrate well with maps yet. Local SEOs are the least disrupted by AI for now.

I trust this data more than the first question above because it asks the respondent to imagine the use case. It is more specific.

---

Cyrus Shepard, Founder at Zyppy “It’s obvious AI has dramatically shifted search behavior – perhaps similar to how the early internet disrupted print. Probably sooner than later, AI search may become so ubiquitous that traditional search seems as quaint as the yellow pages or classified ads.”

---

3. Do people use Google’s AI summaries?

Of course, the line between AI search and traditional search has blurred. Search engines summarize and AI tools search. All models are hybrid models. Traditional SEO is still critical. All systems need to retrieve information.

AI is right there in search results, right at the top. Google’s AI overviews appear in 76% of search results pages. That’s a lot. But how useful are Google’s AI Overviews? What did the respondents say?

Probably because they’re at the top of search results pages, 70% of searchers use AI overviews to get answers. They’re just impossible to miss. Hard not to use them.

That number isn’t growing and some users actually turn AI Overviews off. To do this, click “web” at the top of the search results page. It might be hiding in the “more” dropdown. Google doesn’t make that button easy to find.

4. Which AI chat tools do people use?

Now let’s check the popularity of the platforms.

There are really just six foundational AI platforms (all other AI tools are built on top of these) and they vary widely in popularity. And the relative popularity is changing.

Two takeaways:

• The new AI-native platforms are holding steady. Claude and Perplexity have even seen growth.
• The big-tech platforms have seen some big changes in user adoption. Gemini is way up. Copilot is down.

It’s likely that in the future, Google will become an even stronger player in AI search for reasons we already mentioned. They own the world’s most popular operating system (Android), the most popular browser (Chrome), and the most popular productivity tools (Google Workspace / G Suite) which have a 77% market share in the US.

5. How often do people use AI?

We’ve looked at what tools people use and when. Now we’ll see how often people use AI tools…

AI adoption continues to rise. 72% of respondents use AI at least once a day, up from zero percent three and a half years ago. Most prompts aren’t searches (according to OpenAI, 30% of prompts are the searches). People use AI for all kinds of things.

The data shows that a subset of power users drive a lot of AI use. And that group is growing the fastest. Once people bring AI into their lives, they use it more often for more things, including search: finding answers, getting recommendations and researching buying decisions.

6. Do people trust Google or AI more?

An interesting and important question.

Here we look at trust in Google vs AI chatbots in the context of changing search behavior…

A few key takeaways:

• AI has changed search forever
  Already, 44% of users report a transformation in their experience and expectations. They may be choosing a tool based on their needs (just as we always went straight to Wikipedia sometimes) or they may simply expect one platform to do everything.
• Many users still predict an end to traditional search
  It’s a common prediction. But likely, we’re headed toward a blended experience that combines search and generative AI. It won’t be an either/or outcome. The winning tools both generate answers and retrieve information.
• Trust is eroding, in Google and AI tools
  It’s not surprising to see it in this data. Globally, trust is on the decline. That’s across media, businesses, governments and other organizations.
• Some users say they will never trust AI
  Also not surprising, considering its broad and disruptive impact. Regardless of how useful, it’s a big, sudden change and humans aren’t great at that. Mistrust, skepticism and fear are perfectly natural responses.

What does all this mean for website traffic? Here’s some data from a December 2025 study on organic traffic changes conducted by Graphite using Similarweb data. It shows that traffic is up to both the big sites and small sites. Traffic to the mid-sized publishers has dropped the most. To me, this suggests that search is becoming a more efficient tool for buyers looking for niche providers, possibly because of AI.

Take a minute to imagine the future of discovery

How will users find things a few years from now? What are the biggest implications? Here are a few things I can easily imagine.

• Ads will appear within AI responses for the users of the free versions
• B2B buyers ask AI to recommend partners? “Make me a buyer’s guide for…”
• The importance of a visual internet will change. Websites will increasingly be simple feeds of data into systems that present it in whatever format the user wants. Therefore the use of browsers will decline.
• The future of lead generation and funnels will change. It will look more like this.

What won’t change?

• The importance of informative websites that help visitors and train AIs on the stories of a brand
• The value of differentiated content (original research, contrarian opinions, genuine points of view, sales/marketing aligned content)
• Word of mouth marketing
• Live events and communities (human relationships and hugs)
• Email marketing
• Influencer marketing
• The goal of being found by people who need your insights (content) or help (services)

Channels have changed many times over the last 30 years. But smart brands have always found ways to be discovered, build trust and drive demand. We can do this, people!

Data Summary

I’m putting this here so it’s easier for systems to understand (hope this helps, dearest bots!). Remember, data in charts isn’t extracted by crawlers! Search friendly and AI friendly content is text, not images.

AI Chat tool adoption

• ChatGPT is used regularly by 36% of respondents in both 2025 and 2026 — holding flat as the most-used AI Chat tool
• Google Gemini grew from 29% to 33% regular usage year-over-year, the largest gain of any platform
• Microsoft Copilot declined from 18% to 15% regular usage, the largest drop of any platform
• Claude (Anthropic) grew modestly from 5% to 6% regular usage
• DeepSeek R1 declined from 5% to 3% regular usage

Paid AI Chat adoption

• 41% of respondents pay for no AI Chat tool in 2026, up from 39% in 2025
• Google Gemini is the only major platform gaining paying users, growing from 16% to 18%
• ChatGPT paying users declined slightly from 24% to 23%
• Microsoft Copilot paying users declined from 11% to 8%

AI Chat usage frequency

• 38% of respondents use AI Chat tools many times per day in 2026, up from 34% in 2025
• Daily use (once per day) declined from 28% to 24%, suggesting a shift toward heavier power use
• 31% of respondents use AI Chat tools weekly or less in 2026

How people use AI for research

• 55% of respondents use AI Chat as their primary or frequent research tool in 2026 (almost always + usually combined)
• The share who “usually use AI Chat” for research dropped sharply from 35% to 28% year-over-year
• The share who “usually use Search” for research grew from 22% to 29% — a 7-point swing
• 16% of respondents almost always use a Search Engine for research in both years

AI summarization in Search (e.g. Google AI Overviews)

• 70% of respondents use AI summarization in search results at least occasionally in 2026
• 31% often use AI summarization in search results in 2026, down slightly from 32% in 2025
• The share who have turned off AI summarization grew from 6% to 7% in 2026 — the largest proportional shift in this question
• 23% of respondents ignore AI summarization in search results in 2026, up from 22%

Which tasks people use AI Chat for vs. Search

• AI Chat is preferred over Search for step-by-step instructions by 40% of respondents in both years
• AI Chat gained the most ground on quick factual lookups, growing from 23% to 29% AI Chat preference — a 6-point gain
• Search maintained its strongest advantage for local and business queries, with only 24% preferring AI Chat in 2026, down from 26%
• AI Chat preference for medical information grew from 35% to 36%
• AI Chat preference for vacation planning grew from 34% to 35%
• AI Chat preference for recommendations held at 31% to 32%

Trust and attitudes toward AI Chat vs. Search

• 49% of respondents agree or strongly agree that AI Chat tools will eventually replace traditional Search Engines in 2025, declining to 46% in 2026
• 52% of respondents agreed that they trust Google results more than AI Chat in 2025, declining slightly to 49% in 2026
• 36% of respondents say they are unlikely to ever trust AI Chat tools in 2026, up from 32% in 2025 — the largest attitudinal shift measured
• 44% of respondents say AI Chat tools have changed the way they look for information online in 2026, up from 42% in 2025 — the only trust/attitude metric that improved
• Agreement that respondents are likely to use AI Chat more as tools improve declined from 51% to 49%