<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Alok Ranjan Ojha on Medium]]></title>
        <description><![CDATA[Stories by Alok Ranjan Ojha on Medium]]></description>
        <link>https://medium.com/@alok.ranjan.ojha?source=rss-0749aef07a82------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*XTGsJ6p4FBSc-nev</url>
            <title>Stories by Alok Ranjan Ojha on Medium</title>
            <link>https://medium.com/@alok.ranjan.ojha?source=rss-0749aef07a82------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 16:07:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@alok.ranjan.ojha/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How to Transition from Chatbots to Agents: The 5 Steps That Turn AI into Action]]></title>
            <link>https://medium.com/@alok.ranjan.ojha/how-to-transition-from-chatbots-to-agents-the-5-steps-that-turn-ai-into-action-a58bf962dda8?source=rss-0749aef07a82------2</link>
            <guid isPermaLink="false">https://medium.com/p/a58bf962dda8</guid>
            <category><![CDATA[agentic-ai-framework]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Alok Ranjan Ojha]]></dc:creator>
            <pubDate>Fri, 30 Jan 2026 10:19:51 GMT</pubDate>
            <atom:updated>2026-01-30T10:19:51.693Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>By Alok Ranjan</strong> Enterprise <em>AI Architect &amp; Product Strategist (ISB) </em><a href="https://www.linkedin.com/in/alok-ranjan-ojha/">https://www.linkedin.com/in/alok-ranjan-ojha/</a></p><p>If you’ve spent any time on tech Twitter or LinkedIn lately, you’ve heard the buzzword: <strong>“Agents.”</strong></p><p>Everyone is talking about them, but few people explain what makes an AI <em>act</em> like an agent. We are shifting from the era of <strong>Generative AI</strong> (machines that <em>talk</em> or <em>create</em>) to <strong>Agentic AI</strong> (machines that <em>do</em>).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*BAzo8LnEJpdcBmGHmFB5TQ.png" /></figure><p>Google AI agent framework (captured in the slide above) that perfectly breaks down the anatomy of an AI Agent. It’s very simple, but it holds the secret to how we move from cool demos to software that does our work for us.</p><p>Let’s break down these five pillars <strong>Perception, Reasoning, Planning, Action, and Reflection</strong> and look at how they turn a static language model into an autonomous teammate.</p><p>To make this concrete, let’s use a running example: <strong>Imagine an AI Travel Agent tasked with “Booking a trip to Paris for a vegetarian foodie.”</strong></p><p><strong>1. Perception: The Eyes and Ears</strong></p><p><strong>“Perception is how the agent ingests the messy reality of the world.”</strong></p><p>Before an agent can do anything, it has to understand the context. In the conventional software needed structured data (rows and columns). AI Agents are different, they thrive on unstructured data.</p><p><strong>Perception</strong> is the agent’s ability to take in multi-modal inputs: text, images, audio, or even raw code logs.</p><ul><li><strong>The Example:</strong> Our AI Travel Agent doesn’t just look at a database of flights. It “perceives” your email history to find your frequent flyer number. It “reads” the PDF menu of a restaurant to check if they actually have good vegetarian options, rather than just trusting a “vegetarian-friendly” tag. It might even look at a screenshot of a map you uploaded.</li></ul><p><strong>2. Reasoning: The Brain</strong></p><p><strong>“Data is useless without the logic to process it.”</strong></p><p>Once the agent has perceived the information, it needs to make sense of it. This isn’t just searching for keywords; it’s about context, nuance, and decision-making. This is where the Large Language Model (LLM) shines.</p><p><strong>Reasoning</strong> involves analyzing the perceived data against the user’s goal. It’s the step where the AI says, “Okay, I see the user wants a flight under $600, but I also know they hate early mornings. How do I balance these trade-offs?”</p><ul><li><strong>The Example:</strong> Our Agent sees a $400 flight (great price!) but realizes it has a 50-minute layover in a massive airport (high risk of missing the connection). It uses <strong>reasoning</strong> to decide: <em>“This flight is risky. I should prioritize the slightly more expensive direct flight to ensure they arrive in time for their dinner reservation.”</em></li></ul><p><strong>3. Planning: The Strategy</strong></p><p><strong>“A goal without a plan is just a hallucination.”</strong></p><p>This is the biggest differentiator between LLM models like (ChatGPT, Gemini or Claude…) and a true Agent. If you ask a chatbot to write code, it spits it out. If you ask an Agent to build an app, it <strong>Plans</strong>.</p><p><strong>Planning</strong> is the ability to break a complex, vague goal into small, sequential steps. It creates a “Chain of Thought.”</p><ul><li><strong>The Example:</strong> The Agent creates a checklist:</li></ul><ol><li><em>Search for flights departing Friday evening.</em></li><li><em>Confirm flight availability.</em></li><li><em>Cross-reference hotel locations with top-rated vegetarian restaurants.</em></li><li><em>Draft an itinerary.</em></li><li><em>Verify the total budget.</em></li></ol><p>It knows it cannot book the restaurant <em>before</em> it knows which city area the hotel is in. That is planning.</p><p><strong>4. Action: The Hands</strong></p><p><strong>“Thinking is nice. Doing is better.”</strong></p><p>For a long time, AI was trapped in a text box. It could write an email for you, but you had to copy-paste it and hit send. <strong>Action</strong> is the moment the agent interacts with the outside world via “Tools” or APIs.</p><ul><li><strong>The Example:</strong> Instead of telling you “You should book flight BA304,” the Agent connects to the airline’s API. It authenticates using your credentials, selects the seat, and <strong>actually executes the booking</strong>. It then opens your Calendar app and places a hold on those dates. It has moved from a consultant to an executor.</li></ul><p><strong>5. Reflection: The Conscience (and Teacher)</strong></p><p><strong>“The smartest agents are the ones that learn from their mistakes.”</strong></p><p>This is the final frontier. <strong>Reflection</strong> is the loop where the agent reviews its own work to improve future performance. It evaluates the outcome of its Actions against the original Goal.</p><p>Without reflection, an AI makes the same mistake a thousand times. With reflection, it gets smarter every day.</p><ul><li><strong>The Example:</strong> After the trip is booked, the agent notices a flag: <em>“I booked a hotel that is 5 miles from the city center, and the user had to take a long Uber ride.”</em></li></ul><p>It reflects and stores a memory: <em>“Next time, prioritize proximity to the city center over saving $20 on the nightly rate.”</em> The next time you ask for a trip, it doesn’t make that mistake again.</p><p><strong>Why This Matters</strong></p><p>We are moving away from prompting AI to <strong>delegating</strong> to AI.</p><p>When you look at the diagram above, don’t just see technical terms. See the evolution of work.</p><ul><li><strong>Perception</strong> = Research</li><li><strong>Reasoning</strong> = Analysis</li><li><strong>Planning</strong> = Project Management</li><li><strong>Action</strong> = Execution</li><li><strong>Reflection</strong> = Up skilling</li></ul><p>The “Autonomous entities designed to achieve specific goals” (as the slide defines them) are not sci-fi. They are simply software that has finally learned how to follow the same workflow that high-performing humans do.</p><p>The screen shot of my travel agent</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/936/1*46SCLTKXu32MIm17LMtJ_A.png" /></figure><p><strong>How can I help you regarding AI Agents?</strong></p><p>Would you like me to help you <strong>brainstorm a use case</strong> for an AI Agent in your specific industry? Do you have data security and protection of PII data as your concern? Do you have question about ROI of AI projects? If any of these are your concerns, if you need help just drop a comment.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a58bf962dda8" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>