Putting a prompt together can be nuanced, but it’s necessary. And the way you approach it will determine how valuable the output might be. A good prompting framework goes beyond basic content requests. When done well, it goes into desired tone, specific audience considerations – even multiple perspective generations.

One of the issues for organisations is when they allow staff to use AI without guidance. What I’ve noticed over the past few years is that this builds in inconsistency… and much of that comes down to the prompts that are used, often without strategy.

So, build a prompt library and share it with your team.

Your prompt library should capture successful AI interaction strategies – show them the prompt and the output, then connect it to your writing process and show them the results. And not all of the prompts need to come from you – encourage teams to document and share effective prompts, building a collective knowledge base that improves AI communication capabilities over time. This approach not only builds resources quickly, it puts your team’s skin in the game, which helps with acceptance.

Then implement regular training and skill-sharing sessions focussed on your library. One of the challenges of AI is that continuous learning needs to keep pace with breakneck evolution. To stay ahead of the curve, you need to ensure you’re not just putting out spotfires or relying on staff getting it by osmosis. So show them what it looks like when it’s working well.

Instead show them what you’ll accept, what you want and, most importantly, give them examples of how.

Here’s the challenge I find for corporates and government agencies. I don’t know if you’ve noticed, but recruits are getting younger and younger, and AI is not an option for them. They just use them to enhance their socials, research their holidays or create Lego minifigures of themselves.

This group needs to be shaped and guided, and relying on people to get it right every time isn’t a risk-management strategy.

At the other end of the experience scale are those staff members who are anti-AI. You don’t need to win them over about technology – but you do need to get them on board to the time-saving aspect to AI (if you’ve found one and it’s actually time-saving). That is way better than encourages them to embrace technology they think is a waste of time at best, a portal to SkyNet at worst. (Ask them about SkyNet. They remember the movie …)

Many people use the prompt window from scratch, then re-direct the prompt or refine the prompt based on what it gets back. They think straight into the prompt window.

Here’s another idea which you could build into your process. Develop your prompt in Word, get it right, then paste it into the window. Make it anonymous/redacted/protected/safe outside of AI, then put it into AI. Then let AI suggest ideas and options.

It’s just a small thing, but if you have to be as risk averse as you can, it removes the chance of private, personal, unredacted or classified information going straight from your head into the world. Sure, it adds a minute or two, but it’s actually risk management.

Many people fall into the trap of asking AI to produce the final output, thinking it’s saved them time in creation. But it hasn’t saved any time. It’s just REDISTRIBUTED it. You see, when your AI text doesn’t get the response you need – and it likely won’t – then you now need to spend time redoing that task.

When using AI to generate text, think of it as a sophisticated brainstorming partner that provides a starting point, not the end result. The key is to view AI-generated content as a blank canvas that requires human expertise, nuance, and critical thinking.

AI can help – a lot – with blank page syndrome or writer’s block. By generating an initial draft, you can quickly produce a framework that captures basic information and structure.

But then the next step is key. It has to through your hands. And it’s more than just adding depth, correcting potential inaccuracies.

AI, when it generates text, doesn’t do communication nuances, it doesn’t know your style and it absolutely doesn’t know your audience. It can generate generic text quickly, but cannot replicate the strategic thinking and emotional intelligence that you bring as a human communicator.

Your role is to take the draft and shape it into a compelling, authentic piece of communication. That’s where the time-saving lies.

By treating AI as a first draft generator, you maintain control over the final message while benefiting from its real capability, which is rapid content creation.

AI challenges that. Simply using AI is not problematic in itself, but how you communicate its use determines audience perception. I don’t have advice on the binary of a yes/no disclosure, but you do need to consider it. Why? Because when the question comes – and it’s likely to – you need an answer.

Plus, some audiences and stakeholders will have varying comfort levels with AI technology. Some may be impressed by technological innovation, while others might be skeptical. Others may feel you’re not fulfilling on your promise to connect with them – because you’re cutting corners on that connection. (And remember, your audience, customers and community don’t enjoy the benefit of the time you saved in AI content creation. They don’t even see it… but they see and feel the change in style and distance in what you offer).

So here are some thought starters on how you could approach transparency in AI usage.

• Develop a clear policy that specifies exactly when, why, and how AI is being used in your communication processes. Work out how to implement and share the policy.
• When AI has been involved in content creation, provide specific details about its role. Instead of a vague statement like ‘AI was used’, explain how AI contributed – did it generate initial research, helped draft a document, or assist in data analysis. This helps builds and protect audience trust.
• 
  Be particularly careful in sensitive communication areas like healthcare, legal, or government communications. In these sectors, audiences expect high levels of expertise and personal touch. You might need to explain how AI is a supportive tool that enhances, rather than replaces, human expertise.
• Develop a consistent approach to AI disclosure across all communication channels. Whether it’s a report, social media post, or internal memo, maintain a uniform standard of transparency that reflects your commitment to ethical technology use.

AI was used to generate the thumbnail image associated with this post. It came from Canva’s in-built AI engine. Kinda cool if you’re looking for a difference in your imagery.

Preparing for the future of AI in communication requires walking the tightrope between forward-thinking approach and core communication principles. While technology changes at a rapid rate, people don’t, and we still need to maintain authentic human connection. We also need to have our finger on 1,000 pulses to ensure that we don’t get behind the eight-ball of staff using AI we have never heard of.

It’s a challenge, so here are some ideas to future-proof your comms:

• Show the value of long-term interest in AI that goes beyond the toy nature into the tool/tech side. Dedicated some time or people (it doesn’t have to be much) to explore new tech, assess their impacts (both positively and negatively) and give them criteria to use to evaluate based on those.
• Advise staff that they can put forward recommendations for AI use, but using a business-case approach built on outcome.
• Open the channels with your staff to share AI they’ve seen (anywhere, not just at work) that is impressive, useful or innovative. Widen your research net.
• Establish ongoing dialogue with technological experts, academic researchers, and industry innovators to stay ahead of emerging communication technologies.
• Create partnership programs that provide insights into potential future technological developments in your area. Become the pilot for research that could help position you at the forefront.
• Cultivate a organisational culture that views technological change as an opportunity for growth and innovation. Encourage staff to approach AI not with fear or resistance, but with curiosity, creativity, and a commitment to continuous learning.

Do you know what the original pitch was? It will sell for you while you sleep. That was the direct quote. You won’t need to work. It will work for you … and you can go to the beach or spend your time somewhere else.

How did that go?

Then we were all sold software that would “revolutionise our business and fit seamlessly into our workflow.” Until we installed it and found we now needed to upgrade three other systems, train our staff out of a process that both worked and was embedded, and change our terminology because “the software doesn’t call it that”. (This is a real example with a client of mine with over 3000 staff. They were expected to change their entire vocabulary away from 100 years of industrial tradition to suit this new software).

Fast forward ten years. I had entire financial sectors come to me saying “help us build a profile in this new thing called social media.” We’ve heard this thing is great, so do you know anything about … MySpace? Yeah, it was even pre-Zuck.

What was the original pitch of social media? I was sold on it being a free and easy way of reaching billions of people and make them your potential customers. Note that word there? Easy.

Was that true? Well, it turns out that the people that we thought were friends weren’t actually friends. The people that clicked like didn’t actually like us, and it introduced a range of other issues that we’re trying to now unpick and unwind across society.

We’ve done the same with video – “store your videos on this new YouTube thing” and podcasts: “all you need is a microphone and an idea.” Forgetting of course that if I skip your TV ad while it’s on TV, having it on YouTube doesn’t make it more watchable, while 90% of new podcasts never see Episode 2.

What’s the common mistake that we’ve made? Thinking because something is new it must be used, and the focus is usually on how it will save you time. People have usually approached digital tools as a time saver instead of a producer of quality.

I’m hearing it again with ChatGPT. “I get AI to write me a 250-word blog post on the topic of financial implications of saving for young people and I don’t need to spend an hour applying my 15 years of financial knowledge.” If you head into AI, it will generate everything for you in a way that’s free and easy. Apparently. In other words, we’re trying to solve another communication problem with technology.

We’re making the same mistakes all over again. By all means you can enhance communication and you can improve it. You can shorten the distance between questions and answers, and you can streamline your processes.

But you can’t fix it. And you definitely just can’t replace it.

So if we’re not going to make the same mistakes again, how should we approach AI?

When we did produce great websites they gave easy or quick access, gave broader information, gave us the ability to download useful information and keep it on hand for when we needed it. When we used social media effectively to improve communication, we were able to engage in a two-way conversations with people who were both our customers and also from different markets.

So maybe the best approach with technology and communication is not to think of technology as the solution, but as part of it. So how do you approach AI if you want to use that in that way? Well, it turns out that AI can help shorten the time that you generate that you use to generate ideas. It could be the use AI to help streamline search processes. It could be that you use AI as a way to polish and review text. It could be that you analyze gaps in your content or audience.

Smart businesses will understand that and implement their AI strategies to enhance, backfill or streamline their communication, not try and solve it like it’s an equation.

Unfortunately, most businesses won’t.

I had one like that last week.  I had a client whose web site need reworking … hence my participation in the project.  The web site needed reworking simply because it wasn’t working.  Forget SEO and all that stuff – even though he was high on Google results, his web site wasn’t delivering anything.

I suggested rewriting his web site because of that non-performance.  I suggested starting some sections again and focussing them more on his target audiences.

He didn’t want to do that.  His idea was for me to rearrange the words already there because he had invested a lot of time in SEO and didn’t want to lose his Google ranking.

My argument – which won in the end, thankfully – was that it didn’t matter if he was high on Google’s Holy Grail of Page 1, the web site wasn’t working.

So I started the job as a writer.

And now his web site is starting to transform, in a way that might help drive his business.

Has that happened to you?  As a writer, have you dealt with a project when it felt like you were rearranging the ‘deckchairs on the Titanic’?

On Australia Day, I read several online articles about Australians who had been given gongs, in recognition of great work done for the community, industry or Australiankind. All positive stuff and heartwarming to read.

I then did what many people do when reading about the long list of Australians given awards … I wondered if any of them were people I knew or worked with. So I headed over to the Government web site which lists the awards.

And that’s where the issues started to raise their heads. (As an aside: this isn’t a direct criticism of the determined footsoldiers running this web site – they’re often guided by misdirection. What my experience did do was just highlight for me issues that I see in Government web sites again, and again, and again.)

I went to the web site – on Australia Day – to read the list, but it wasn’t there.  It was going to be added to the site ‘in the near future’.

Issue #1: your content needs to reflect your audience and their needs – both they what and, importantly, when they want it.

I’ve worked with some departments who’ve uploaded content when THEY were ready. That’s not the point. People will visit you once then, if they can’t find what they’re looking for, go away and never return.

The other thing I found on this web site was terminology telling how HOW to access information if and when it was made available. All very process-focussed – but still not answering my informational need.  There were lines and lines about how it might work.

Issue #2: don’t talk about process – especially when people are told they can’t have what they want.

People aren’t interested in process – the idea of ‘click here’, ‘visit this section later’ etc. They want content – then they’ve moved on.

So did I go back to find out the answers to my questions? No. As a consumer I’d moved on by then and, when a colleague raised these awards, but first thought – and comment – was how hard the web site was to use.  January 27 was no longer Australia Day, so I’d moved on.

That’s how people respond to web sites that don’t help them. They move on, filing away mentally that your web site is hard work.  And, sadly based on my experience of working with many Government Departments, in Government that can happen a lot.

Issue #3: the thing people are interested in is what your web site can provide, not that you have a web site.

This is 2025. Australian web sites began 30 years ago, and they’re now an expected part of business. We – as your audience – don’t visit your web site because it’s your web site. We go there because it contains the answer we need, the solution to our problem or to get the contact name or number to have a conversation.

I still hear this when working with Government departments in my workshops. The key audience is sometimes mistakenly thought of as “The Minister”. (Again, this isn’t a shot at the footsoldiers who clearly don’t agree. This is about the hierarchy.) But that type of response produces pretty brag books that don’t engage – and in turn chip away at the veneer of the department’s (or Minister’s) reputation.

When it comes to change, people don’t like it – even and especially they complain about the current situation. So how do you position change with people?

It’s about addressing the right issue. People are comfortable with the current situation because it’s known, and it’s their process. They may not love it, but they know it.

So if you want to introduce change, you need to get them to buy into the change that you are suggesting.

Just say I try to convince you in the traditional way, which is to try and make it exciting. This is the guaranteed fail way of communicating change. I’m addressing the wrong issue.

You see, if you’re reticent to change because you’re comfortable in what you’re doing, telling you it’s new and exciting doesn’t work, because your problem isn’t that you’re bored. In fact, you don’t actually think you’ve got a problem.

To get you to buy into the change, or whenever you want to change people’s behaviour, I need to give you a rear-view mirror look at the result of the change, and it has to be tangible.

Say I’ve got a new software system (and this mistake is made most with this type of communication). I could talk about our exciting and new Cloud-based ‘whatever’ but I need to focus on the time saved, the easier process, the benefits to me. I’m now positioning the change in terms of output, not process.