<?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 Ashish Mohan on Medium]]></title>
        <description><![CDATA[Stories by Ashish Mohan on Medium]]></description>
        <link>https://medium.com/@ashishmohantehd?source=rss-bf8951c7d613------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*1dd80xszcRnXgBzJw5rWFA@2x.jpeg</url>
            <title>Stories by Ashish Mohan on Medium</title>
            <link>https://medium.com/@ashishmohantehd?source=rss-bf8951c7d613------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 23:10:54 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ashishmohantehd/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[Nobody Told You That AI Agents Are Now Your Competition – Not ChatGPT]]></title>
            <link>https://medium.com/@ashishmohantehd/nobody-told-you-that-ai-agents-are-now-your-competition-not-chatgpt-844c12b8295d?source=rss-bf8951c7d613------2</link>
            <guid isPermaLink="false">https://medium.com/p/844c12b8295d</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Ashish Mohan]]></dc:creator>
            <pubDate>Fri, 29 May 2026 03:10:22 GMT</pubDate>
            <atom:updated>2026-05-29T03:10:22.980Z</atom:updated>
            <content:encoded><![CDATA[<p><em>The enterprise shift happening this quarter is different from anything in the past three years. Here’s what it means for Business Analysts, PMs, and every knowledge worker who hasn’t updated their mental model yet.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/392/1*2ztvWamppHzKTQWkF9BHOQ@2x.jpeg" /></figure><p>Six months ago, the threat felt theoretical.</p><p>A colleague forwarded me a Slack message from a Head of Digital Transformation at a major US retailer. It read: “We just cut our BA intake by 60% for Q3. Not because the work went away. Because our agents do the first three weeks of it automatically now.”</p><p>He sent it half as a joke. I didn’t laugh.</p><p>Because today, Thursday May 28, 2026, that message isn’t unusual anymore. It’s a template.</p><p>This is the shift that changes everything – and most professionals in the analysis and strategy space are still framing it the wrong way.</p><p><strong>The Conversation Nobody Is Having Correctly</strong></p><p>For the past three years, the conversation in business analysis and product circles has been: <em>“Will AI replace BAs?”</em></p><p>That was the wrong question. It framed AI as a single tool like Excel or JIRA – that either threatens your job or doesn’t.</p><p>The right question, the one that actually maps to what’s happening in enterprise right now, is this: <strong>Will you be the person who orchestrates AI agents, or will you be the person those agents were hired to replace?</strong></p><p>Those are not the same question. The first is about a tool. The second is about a role.</p><p>Gartner now predicts that 40% of enterprises will use task specific AI agents by the end of 2026, up from less than 5% in 2025 and McKinsey data shows 62% of organizations are already experimenting with them.  These aren’t chatbots answering FAQ questions. These are autonomous systems that receive a business objective, decompose it into subtasks, select their own tools, and execute end-to-end workflows – without a human in the loop for each step.</p><p>A requirements-gathering agent. A process gap analysis agent. A stakeholder interview synthesis agent.</p><p>Sound familiar? It should. That’s a junior BA’s job description.</p><p><strong>What the Data Actually Shows About Where We Are</strong></p><p>Let me ground this in what’s real today, not the hype from a vendor keynote.</p><p>According to the U.S. Census Bureau’s Business Trends and Outlook Survey which tracked data from December 2025 through early May 2026 overall AI usage among US businesses has held between 17% and 20%, with 20 – 23% of businesses expecting to adopt AI in the next six months. </p><p>That number sounds modest until you understand the distribution. Small businesses skew that average down dramatically. In large enterprises the Fortune 1000, the consulting firms, the financial institutions where most senior BAs work adoption is far higher and accelerating faster.</p><p>Deloitte’s 2026 State of AI in the Enterprise report found that worker access to AI rose 50% in 2025 alone, and the number of companies with 40% or more of their AI projects in production is set to double within six months. </p><p>That last line deserves a pause. Double in six months. Not double by 2028. By November.</p><p>The IIBA’s own 2025 Global State of Business Analysis Report confirmed something practitioners have been feeling for a year: 76% of BAs now report that business analysis is playing a larger role in strategic decision-making even as the traditional, output-focused work is being automated beneath them. </p><p>This is the paradox that defines the moment. The role is becoming more strategic at the exact same time the entry-level work is being automated away. The ladder is losing its bottom rungs – and the top is rising faster than most people can climb.</p><p><strong>The Three-Layer Model: Where You Actually Stand</strong></p><p>I’ve started thinking about enterprise BA and PM work in three layers, because it clarifies what’s at risk and what isn’t.</p><p>Layer 1 – Execution work. Documentation, meeting summaries, first-draft user stories, gap analysis scaffolding, process mapping templates, SQL queries, status reports. This is the layer AI agents are consuming right now. If 60 – 70% of your current week lives here, you have a 24-month window to move.</p><p>Layer 2 – Synthesis work. Taking multiple inputs – data, stakeholder perspectives, organizational history, market context and generating a coherent, opinionated point of view that a decision-maker can act on. AI does a competent version of this. A skilled human does a better version with context AI will never have access to. This layer is contested. The best professionals are winning it by a large margin. The average ones are losing ground quietly.</p><p>Layer 3 – Judgment work. Problem framing. Stakeholder trust. Political navigation. Knowing which requirements to push back on and why. Understanding that the reason the project is stuck isn’t the backlog – it’s that two VPs haven’t been in the same room since the reorg. AI cannot touch this layer. It doesn’t even know it exists.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/392/1*2v8XnMuOo930e7YkNO7ZgA@2x.jpeg" /></figure><p>The professionals who are thriving right now have deliberately moved as much of their identity – and their calendar – into Layer 3 as possible. The ones who are anxious are still delivering Layer 1 work manually and wondering why they feel replaceable.</p><p><strong>The Agentic Workflow: What It Actually Looks Like in Enterprise Today</strong></p><p>Here’s a concrete example of what an AI agent deployment looks like in a mid-large enterprise as of this quarter because the abstraction often obscures the reality.</p><p>A financial services firm rolls out an agentic requirements intake system. When a business unit submits a new initiative request, the agent automatically: pulls the relevant historical project data, cross-references it against current system architecture, generates a preliminary scope document, flags likely dependencies and risks, and routes a structured brief to the BA assigned to the initiative.</p><p>By the time a human BA touches the project, three to four days of traditional intake work is already done.</p><p>The BA’s job is not to redo that work. It is to do what the agent cannot: validate the framing against unstated organizational priorities, build the stakeholder relationships that make execution possible, and catch the thing the agent got wrong – because it always gets at least one thing wrong, and it doesn’t know it.</p><p>This is precisely where augmented analytics is failing in organizations with immature data governance: when users can’t verify the reasoning behind AI-generated recommendations, they reject them and revert to manual analysis. A B2B SaaS company abandoned a $300K analytics implementation because the system couldn’t explain its own methodology to a senior executive. </p><p>That explanation gap between what AI produces and what a human leader will trust is not closing. It is a permanent feature of how decisions get made in organizations with real stakes. The BA or PM who can bridge that gap is not less relevant. They are load-bearing.</p><p><strong>What Leaders Are Getting Wrong This Quarter</strong></p><p>I’ll be blunt about what I’m observing in enterprise right now.</p><p><strong>Deploying agents without governance.</strong> Organizations are standing up agentic workflows requirements intake, analysis, documentation without defining who is accountable when the agent produces something wrong and a team builds on it for two weeks. This is not a technology problem. It is an accountability architecture problem. It is exactly the kind of problem a senior BA should be designing the solution for and most aren’t being invited to that conversation yet.</p><p><strong>Confusing speed with quality.</strong> Agentic AI produces faster artifacts. Faster is not better by default. The organizations seeing the most value are the ones that defined quality standards first and then measured AI output against them. The ones chasing speed without quality frameworks are producing documented garbage at scale.</p><p><strong>Eliminating the junior pipeline.</strong> The same firms cutting junior BA intake to save cost are eliminating the training ground that produces senior analysts in three to five years. This is a decision that looks efficient in 2026 and catastrophic in 2029.</p><p><strong>What I Would Do If I Were Starting My BA Career Today</strong></p><p><em>Three things. Right now. This week.</em></p><p><strong>Own one agent workflow end-to-end.</strong> Find a repetitive, document-heavy deliverable in your current role and build a full agentic workflow around it – prompt chain, validation step, output template. Not to save time (though you will). To understand what agents can and cannot do from the inside. That understanding is the basis of every smart conversation you’ll have for the next five years.</p><p><strong>Move your visible value to Layer 3.</strong> In your next stakeholder meeting, your next project review, your next 1:1 with your manager lead with a judgment call, a framing observation, a political read. Make the layer AI cannot access the layer people associate with you.</p><p><strong>Get fluent in agent failure modes. </strong>Study the cases where enterprise AI deployments have failed data quality issues, explainability gaps, stakeholder rejection. This is your consulting edge. The organizations deploying agents in 2026 need someone who can diagnose why the agent is wrong. That person needs to understand both the business context and the AI limitation simultaneously. Almost nobody can do both. Yet.</p><p><strong>Prediction: What Q4 2026 Looks Like</strong></p><p>By the end of this year, 40% of US enterprises will have task-specific AI agents in production deployment.  That means by December, nearly half of the organizations where analysts and product managers work will have autonomous systems executing work that was on someone’s job description eighteen months ago.</p><p>The professionals who will be in the strongest position in Q4 2026 are not the ones who used AI the most. They’re the ones who understood what AI changed about their specific value – and adjusted accordingly.</p><p>The BA who can walk into a room in December and say: “Here’s what our agents produced. Here’s where they’re wrong. Here’s the decision we should actually make, and here’s why” – that person is not threatened by agentic AI.</p><p>That person is the reason the agents are worth deploying.</p><p>Where are you in this shift are you building toward Layer 3, or still fighting to protect Layer 1? I’m genuinely curious what’s happening in your organization right now. Drop it in the comments the most useful signal I get comes from practitioners, not research reports.</p><p><em>Follow for weekly analysis on enterprise strategy, AI adoption, and the evolving role of Business Analysis in US organizations.</em></p><p>Business Analysis · Artificial Intelligence · Product Management · Enterprise Technology · Career Development</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=844c12b8295d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Meeting That Changed How I Think About AI and Business Analysis]]></title>
            <link>https://medium.com/@ashishmohantehd/the-meeting-that-changed-how-i-think-about-ai-and-business-analysis-22f4c504d420?source=rss-bf8951c7d613------2</link>
            <guid isPermaLink="false">https://medium.com/p/22f4c504d420</guid>
            <category><![CDATA[business-analysis]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[career-development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Ashish Mohan]]></dc:creator>
            <pubDate>Sat, 23 May 2026 23:50:10 GMT</pubDate>
            <atom:updated>2026-05-23T23:50:10.814Z</atom:updated>
            <content:encoded><![CDATA[<p><em>One conversation with a executives revealed the biggest misconception enterprises have about AI replacing analysts and what smart professionals are doing about it right now.</em></p><p>I almost lost a client last month.</p><p>Not because of bad work. Not because of missed deadlines. Because I was too slow.</p><p>The executives had asked for a competitive landscape analysis market sizing, operational gaps, three strategic options with financial modeling. Standard BA deliverable. I quoted ten business days. She paused, looked at me across the table, and said something I haven’t stopped thinking about:</p><p>“Our AI team said they could have a first draft by tomorrow morning.”</p><p>She wasn’t threatening me. She was genuinely curious what she was paying for.</p><p>That question – what are you paying for? is the most important question in Business Analysis right now. And most professionals don’t have a clean answer.</p><p><strong>The Shift Happened Faster Than Anyone Predicted</strong></p><p>In January 2023, most enterprise teams were still debating whether AI tools were “ready for production.” By January 2025, Microsoft reported that over 70% of Fortune 500 companies had active Copilot deployments. Gartner’s latest data shows that knowledge worker productivity in analysis-heavy roles has improved 30 – 45% at organizations with mature AI adoption.</p><p>That’s not a prediction. That’s already in the rearview mirror.</p><p>The BAs and PMs who treated 2023 and 2024 as “wait and see” years are now playing catch-up against colleagues who spent those two years building workflows, prompt libraries, and AI-native habits. The gap is visible in output quality, in delivery speed, and increasingly in compensation.</p><p>According to LinkedIn’s 2025 Jobs on the Rise report, roles combining “business analysis” with AI fluency are commanding 28 – 40% salary premiums over traditional BA positions with equivalent experience. That spread is widening, not stabilizing.</p><p><strong>What AI Is Actually Good At (And What It Isn’t)</strong></p><p>Here is the honest version not the vendor pitch, not the tech-pessimist backlash. Just what I observe across enterprise engagements.</p><p>AI is genuinely transformative at synthesis and first-draft generation. Feed it a 90-minute stakeholder transcript and a structured prompt, and it will produce a usable requirements brief in four minutes. That used to take half a day. Feed it your current-state process documentation and your future-state objectives, and it will identify gaps, redundancies, and risks in a form that would have taken a junior analyst two days to assemble.</p><p>This is real. This is happening. And it changes the economics of junior BA work significantly.</p><p>But here is what AI cannot do and this is not a temporary limitation waiting on the next model release:</p><p>AI cannot read a room. It cannot sense that the VP of Finance is nodding along but has no intention of funding the initiative. It cannot detect that the stated requirement (“we need a dashboard”) is a political deflection from the real problem (two departments don’t share data because their leaders haven’t spoken directly in eight months). It cannot build the kind of trust where a stakeholder tells you the thing they would never put in an email.</p><p>That interpretive, relational, politically-aware layer of Business Analysis is not augmented by AI. It is made more valuable by AI because it becomes the scarce, differentiating skill in a world where the documentation work is commoditized.</p><p><strong>The Mistake Most Leaders Are Making Right Now</strong></p><p>Across the organizations I work with, the most common AI mistake is not a technology failure. It is a governance failure.</p><p>Leaders are deploying AI tools – Copilot, Claude, Gemini, whatever the procurement team approved – without defining what decisions AI is allowed to influence and which ones require human judgment. There is no policy. There is no framework. There is just a rollout email and an assumption that professionals will self-regulate.</p><p>They won’t. Not because they’re irresponsible, but because nobody has given them a reason to slow down. AI generated analysis feels authoritative. It is formatted beautifully. It cites things. And it is sometimes wrong in ways that are very hard to detect without deep domain expertise.</p><p>The organizations that are winning with enterprise AI have a human-in-the-loop standard: every AI-generated artifact above a certain business impact threshold requires a qualified professional to validate it before it influences a decision. That’s not anti-AI. That’s sound risk management – and it’s exactly the kind of governance framework a senior BA is positioned to design and own.</p><p><strong>What I Told the executives</strong></p><p>I went back to executives conversation with a different answer than I would have given a year ago.</p><p>I told them: their AI team could have a draft by morning. I could have a decision by end of week.</p><p>The draft would give her information. What I deliver is judgment about which options are politically viable given her organization’s structure, which financial assumptions are optimistic versus defensible, and which strategic path is most likely to survive contact with her board.</p><p>She gave me the project.</p><p>Not because I was defensive about AI. Because I was clear about what I offer that AI doesn’t.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/584/1*Myq8ljo6qtmuLD_n-DBDzQ@2x.jpeg" /></figure><p><em>Image: The meeting that changed how I think about AI in business analysis – and why human judgment still wins.</em></p><p><strong>What You Should Do This Week</strong></p><p><em>Three things. Not ten. Three.</em></p><p><em>First</em>, take one recurring deliverable you produce – a meeting summary, a gap analysis, a requirements brief – and build a repeatable AI prompt that produces an 80% draft. Refine it until it reliably works. That’s your first workflow asset.</p><p><em>Second</em>, identify the one capability you have that AI demonstrably cannot replicate in your specific organizational context. Political knowledge. Domain depth. A specific stakeholder relationship. Get very clear on it, because that is your professional moat.</p><p><em>Third</em>, update how you describe your work to your manager, on your LinkedIn, in your next stakeholder meeting. You are not a documentation professional. You are a strategic analyst who uses AI to eliminate low-value work and focus judgment where it matters most.</p><p><strong>The Bottom Line</strong></p><p>The executives question – what are we paying for? is the question every BA and PM should be asking themselves before their clients ask it for them.</p><p>The answer, in 2025, is not requirements documents. It is not process maps. It is not user stories.</p><p>The answer is judgment, trust, and the ability to turn ambiguous organizational reality into decisions that stick.</p><p>AI makes that work faster to prepare for and harder to fake. That is, depending on where you sit, either the best or worst thing that has happened to this profession.</p><p>I think it’s the best. But only if you decide it is and then go build something to prove it.</p><p><strong>Which part of your BA or PM work do you think AI genuinely cannot replace? </strong>I’m collecting perspectives from practitioners across industries, drop your answer in the comments.</p><p><em>Follow for weekly analysis on enterprise strategy, AI, and the future of knowledge work.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=22f4c504d420" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>