<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>Shravan Weblog</title>
	<atom:link href="https://shravan.dev/feed/" rel="self" type="application/rss+xml" />
	<link>https://shravan.dev</link>
	<description></description>
	<lastBuildDate>Thu, 19 Mar 2026 19:53:35 +0000</lastBuildDate>
	<language>en-US</language>
	<sy:updatePeriod>
	hourly	</sy:updatePeriod>
	<sy:updateFrequency>
	1	</sy:updateFrequency>
	<generator>https://wordpress.org/?v=7.0</generator>

<image>
	<url>https://i0.wp.com/shravan.dev/wp-content/uploads/cropped-site-icon-150x150-1-1.webp?fit=32%2C32&#038;ssl=1</url>
	<title>Shravan Weblog</title>
	<link>https://shravan.dev</link>
	<width>32</width>
	<height>32</height>
</image> 
<site xmlns="com-wordpress:feed-additions:1">31926741</site>	<item>
		<title>Breaking Free from the In-House Enterprise AI Platform Trap: A Practitioner&#8217;s Playbook</title>
		<link>https://shravan.dev/breaking-free-from-the-in-house-enterprise-ai-platform-trap-a-practitioners-playbook/</link>
					<comments>https://shravan.dev/breaking-free-from-the-in-house-enterprise-ai-platform-trap-a-practitioners-playbook/#respond</comments>
		
		<dc:creator><![CDATA[Shravan Kumar Kasagoni]]></dc:creator>
		<pubDate>Thu, 19 Mar 2026 19:11:51 +0000</pubDate>
				<category><![CDATA[ARTIFICIAL INTELLIGENCE]]></category>
		<category><![CDATA[AUTOMATION]]></category>
		<category><![CDATA[GENERATIVE AI]]></category>
		<guid isPermaLink="false">https://shravan.dev/?p=9348</guid>

					<description><![CDATA[From Diagnosis to Prescription In Part 1 of this series, the diagnosis was laid bare: nine interlocking traps that cause in-house enterprise AI platforms to consume budgets, produce polished demos, and deliver zero measurable business impact. The data was unambiguous: internal AI builds succeed at one-third the rate of purchased solutions (MIT NANDA, 2025), 95% of enterprise AI pilots fail to move the P&#38;L, and not one of 598 published enterprise AI case studies contained rigorous evidence of business impact (Applied AI, 2025). But a diagnosis without a prescription is just complaining. This second part is the practitioner’s playbook: six concrete strategies for enterprises that want AI to deliver outcomes rather than architecture diagrams. These are not theoretical frameworks. They are drawn from direct experience leading automation at scale in a regulated, global enterprise, and are validated against the research that exposed the traps in Part 1. The unifying principle is simple: own the domain knowledge, rent the AI infrastructure. Strategy 1: Invest in Process Intelligence, Not Model Infrastructure The competitive moat for any enterprise is not the ability to run an LLM. That capability is fully commoditised. OpenAI, Anthropic, Google, and a dozen open-source alternatives provide model access at commodity prices. Wrapping one of these models in a corporate UI adds no defensible value.The moat is understanding your own processes deeply enough to know exactly where AI creates measurable impact: which handoffs introduce delay, which decisions could be automated, where rework occurs, and what the cost of each inefficiency is. Process mining, workflow analysis, and operational data quality are the foundations that determine whether any AI initiative, bought or built, will succeed. This is precisely the gap that current tools leave unaddressed. Descriptive process mining tells you what happened. The real value lies in prescriptive capabilities, recommending specific process...]]></description>
		
					<wfw:commentRss>https://shravan.dev/breaking-free-from-the-in-house-enterprise-ai-platform-trap-a-practitioners-playbook/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
		<post-id xmlns="com-wordpress:feed-additions:1">9348</post-id>	</item>
		<item>
		<title>The In-House Enterprise AI Platform Paradox</title>
		<link>https://shravan.dev/the-in-house-enterprise-ai-platform-paradox/</link>
					<comments>https://shravan.dev/the-in-house-enterprise-ai-platform-paradox/#respond</comments>
		
		<dc:creator><![CDATA[Shravan Kumar Kasagoni]]></dc:creator>
		<pubDate>Sun, 15 Mar 2026 09:26:17 +0000</pubDate>
				<category><![CDATA[ARTIFICIAL INTELLIGENCE]]></category>
		<category><![CDATA[AUTOMATION]]></category>
		<category><![CDATA[GENERATIVE AI]]></category>
		<guid isPermaLink="false">https://shravan.dev/?p=9277</guid>

					<description><![CDATA[Over the last couple of years, many companies have pushed to build their own in-house enterprise AI platforms. The logic: gain control, protect IP, and outpace competitors. CTOs and CIOs feel pressure from boards to present a clear AI strategy, and “building a platform” is often seen as the most ambitious answer. But this approach rarely delivers. In-house AI platforms burn resources, drain budgets, and produce impressive demos that rarely achieve real business outcomes. This issue is not merely theoretical. Recent data provide compelling evidence of its severity. The Numbers That Should Alarm Every Enterprise Leader 95% of enterprise AI pilots fail to achieve rapid revenue acceleration.— MIT NANDA, “The GenAI Divide: State of AI in Business 2025” Internal AI builds succeed only 22% of the time, versus 67% for purchased solutions from specialized vendors.— MIT NANDA, 2025 Over 40% of agentic AI projects will be cancelled by end of 2027 due to escalating costs, unclear business value, or inadequate risk controls.— Gartner, June 2025 The share of companies abandoning most AI initiatives jumped from 17% in 2024 to 42% in 2025.— S&#38;P Global Market Intelligence, 2025 For every 33 AI pilots a company launches, only 4 make it to production — an 88% failure rate.— IDC Research, 2025 These findings reflect consensus among institutions such as MIT, Gartner, Deloitte, McKinsey, IDC, and S&#38;P Global, all of which conclude that current internal enterprise AI approaches are fundamentally flawed. From observing this pattern across industries, nine recurring traps explain why in-house enterprise AI platforms struggle. Trap #1: The Glorified Wrapper Problem Most in-house &#8220;AI platforms&#8221; are just wrappers on foundation models, like GPT or Claude, equipped with an RAG pipeline. A corporate interface and SSO are implemented and shipped as internal products, celebrated by teams and executives. Meanwhile, 90% of employees...]]></description>
		
					<wfw:commentRss>https://shravan.dev/the-in-house-enterprise-ai-platform-paradox/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
		<post-id xmlns="com-wordpress:feed-additions:1">9277</post-id>	</item>
		<item>
		<title>The Evolution of Automation: From Macros to Intelligent Automation</title>
		<link>https://shravan.dev/the-evolution-of-automation-from-macros-to-intelligent-automation/</link>
					<comments>https://shravan.dev/the-evolution-of-automation-from-macros-to-intelligent-automation/#respond</comments>
		
		<dc:creator><![CDATA[Shravan Kumar Kasagoni]]></dc:creator>
		<pubDate>Sat, 05 Apr 2025 09:35:00 +0000</pubDate>
				<category><![CDATA[ARTIFICIAL INTELLIGENCE]]></category>
		<category><![CDATA[AUTOMATION]]></category>
		<category><![CDATA[COGNITIVE AUTOMATION]]></category>
		<category><![CDATA[PROCESS AUTOMATION]]></category>
		<category><![CDATA[Automation Trends]]></category>
		<category><![CDATA[Business Process Automation]]></category>
		<category><![CDATA[Cognitive Automation]]></category>
		<category><![CDATA[Digital Transformation]]></category>
		<category><![CDATA[Process Automation]]></category>
		<category><![CDATA[Robotic Process Automation]]></category>
		<guid isPermaLink="false">https://shravan.dev/?p=8307</guid>

					<description><![CDATA[The journey of process automation is a testament to human ingenuity and our relentless pursuit of efficiency. From the early days of macros and scripting to the sophisticated realms of intelligent automation, this evolution reflects our growing capability to harness technology to transform our work. This blog post delves into each stage of this evolution, providing insights and examples to illustrate the transformative impact of automation technologies. Macros and Scripting:The earliest form of process automation can be traced back to using macros and scripting. Macros are sets of instructions that automate repetitive tasks within software applications. They are the forerunners of automation, enabling users to save time on routine tasks such as data entry, formatting, and simple calculations. For example, in Microsoft Excel, macros can automate complex tasks, like applying specific formatting across multiple datasets, consolidating data from various sheets, or generating standardized reports with a single click. Scripting involves writing short programs or scripts to automate tasks across applications. Scripts can range from simple batch files that automate file management tasks in operating systems to more complex scripts written in languages like Python or PowerShell that can manipulate data, automate server setups, or scrape information from the web. Robotic Process Automation (RPA):As technology advanced, the limitations of macros and basic scripting led to the emergence of Robotic Process Automation (RPA). RPA takes automation several steps further by using software robots, or &#8220;bots,&#8221; to emulate and integrate the actions of a human interacting within digital systems to execute a business process. RPA bots can capture data, trigger responses, initiate new actions, and communicate with other systems autonomously. For instance, RPA is widely used in the banking and finance sector to open accounts, process loans, and handle customer queries. A notable example includes invoice processing automation, where RPA tools extract relevant...]]></description>
		
					<wfw:commentRss>https://shravan.dev/the-evolution-of-automation-from-macros-to-intelligent-automation/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
		<post-id xmlns="com-wordpress:feed-additions:1">8307</post-id>	</item>
		<item>
		<title>Common Mistakes Organizations Make in Automation</title>
		<link>https://shravan.dev/common-mistakes-organizations-make-in-automation/</link>
					<comments>https://shravan.dev/common-mistakes-organizations-make-in-automation/#respond</comments>
		
		<dc:creator><![CDATA[Shravan Kumar Kasagoni]]></dc:creator>
		<pubDate>Tue, 25 Mar 2025 13:14:00 +0000</pubDate>
				<category><![CDATA[ARTIFICIAL INTELLIGENCE]]></category>
		<category><![CDATA[AUTOMATION]]></category>
		<category><![CDATA[COGNITIVE AUTOMATION]]></category>
		<category><![CDATA[PROCESS AUTOMATION]]></category>
		<category><![CDATA[Automation]]></category>
		<category><![CDATA[Business Process Automation]]></category>
		<category><![CDATA[Digital Transformation]]></category>
		<category><![CDATA[Process Automation]]></category>
		<guid isPermaLink="false">https://shravan.dev/?p=8283</guid>

					<description><![CDATA[In the digital transformation era, automation has become a cornerstone for organizations looking to improve efficiency, reduce costs, and enhance productivity. However, the path to successful automation is riddled with potential pitfalls. Understanding these common mistakes can help organizations navigate their automation journey more effectively. 1. Lack of Clear StrategyOverlooking the Big Picture: Organizations often rush into automation without having a clear strategy or defined objectives. They tend to concentrate on automating individual tasks without considering the overall business process. This piecemeal approach may result in suboptimal outcomes and missed opportunities for more impactful automation. 2. Underestimating the ComplexityMisjudging Integration Challenges: Integrating various systems and technologies for automation can be complex, leading to delays and increased costs. A thorough analysis of existing IT infrastructure and a well-planned integration strategy are crucial. 3. Ignoring Employee InvolvementOverlooking the Human Factor: Successful automation isn&#8217;t just about technology; it&#8217;s also about people. Failing to involve employees in the automation process can lead to resistance and reduced effectiveness. It&#8217;s essential to communicate the benefits, provide training, and address employees&#8217; concerns. 4. Insufficient Testing and ValidationSkipping Rigorous Testing: Organizations often neglect comprehensive testing and validation in their rush to implement automation. This oversight can result in system failures and unanticipated problems. It is critical to ensure that automated processes work as intended. 5. Over-AutomationLosing the Human Touch: Automation can improve efficiency, but overdoing it may cause a loss of personal touch and flexibility. Balancing automated and human processes is crucial. 6. Neglecting Continuous ImprovementTreating Automation as a One-Time Project: Automation is an ongoing journey, not a one-time project, and requires regular reviews and updates to align with business goals as technologies and business needs evolve. 7. Inadequate Data Security MeasuresUnderestimating Security Risks: With automation, data security becomes even more critical. Organizations often overlook the increased security...]]></description>
		
					<wfw:commentRss>https://shravan.dev/common-mistakes-organizations-make-in-automation/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
		<post-id xmlns="com-wordpress:feed-additions:1">8283</post-id>	</item>
		<item>
		<title>Setting Up TensorFlow in RStudio on macOS: A Step-by-Step Guide</title>
		<link>https://shravan.dev/setting-up-tensorflow-in-rstudio-on-macos-a-step-by-step-guide/</link>
					<comments>https://shravan.dev/setting-up-tensorflow-in-rstudio-on-macos-a-step-by-step-guide/#respond</comments>
		
		<dc:creator><![CDATA[Shravan Kumar Kasagoni]]></dc:creator>
		<pubDate>Wed, 05 Feb 2025 20:20:51 +0000</pubDate>
				<category><![CDATA[PRODUCTIVITY]]></category>
		<category><![CDATA[TIPS & TRICKS]]></category>
		<category><![CDATA[RSTUDIO]]></category>
		<guid isPermaLink="false">https://shravan.dev/?p=8474</guid>

					<description><![CDATA[1) Install Anaconda Download Anaconda: Visit the Anaconda Distribution page and download the installer suitable for macOS. Install Anaconda: Run the downloaded installer and follow the on-screen instructions to complete the installation. 2) Open Terminal 3) Create a Conda Environment 4) Activate the Environment 5) Install Python 6) Install tensorflow 7) Configure R to Use the Conda Environment Inform R about the specific Python environment to ensure seamless integration. Edit the .Renviron File: This file is typically located in your home directory. If it doesn&#8217;t exist, you can create it. 8) Verify TensorFlow Installation 9) Go to RStudio and Restart the R Session To ensure RStudio recognizes the new environment, Navigate to Session -&#62; Restart R within RStudio. 10) Initialize and Test TensorFlow in R Run the following R script to verify the setup:: After following these steps, you should have a functional TensorFlow setup in RStudio on your macOS system. Ensure all paths are correctly specified, and you use compatible versions of Python, TensorFlow, and R packages.]]></description>
		
					<wfw:commentRss>https://shravan.dev/setting-up-tensorflow-in-rstudio-on-macos-a-step-by-step-guide/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
		<post-id xmlns="com-wordpress:feed-additions:1">8474</post-id>	</item>
	</channel>
</rss>
