# Anubhav Pateriya > Global Head of Consumer Consulting at Wipro and ex-BCG Partner on AI, judgment, and the self-driving consumer enterprise — the Destination Problem. Public Ghost content for AI and LLM tooling. This file includes a bounded export of public pages first, then recent public posts. Append `.md` to any post or page URL to get the content in Markdown (for example, `/example-post.md`). ## Pages ### About URL: https://www.anubhavpateriya.com/about/ Last updated: 2026-07-10T23:52:52.000Z ## I help consumer leaders turn AI from slideware into P&L. For twenty years I've worked every layer where transformation actually succeeds or dies — as the strategist who sets the direction, the operator who owns the number, and now the leader building the practice. I'm Managing Partner and Global Head of Consumer Consulting at Wipro, leading strategy and commercial transformation for CPG, retail, travel and hospitality clients. Before Wipro I was a Partner at BCG, ran a full US P&L as a Senior Partner at Genpact, and built commercial and digital capability at Strategy& and Accenture. Across the US, Europe and Asia the through-line has been the same: making AI and digital show up in margin, growth and the operating model — not just the roadmap. ## The Destination Problem The consumer enterprise is becoming self-driving — demand sensed rather than surveyed, supply chains that negotiate themselves, pricing and innovation encoded into agents. As AI takes over how the business runs, execution stops being the constraint, and advantage shifts to the choices only leadership can make. AI is solving the driving problem. Consumer leaders now face the destination problem. Execution is becoming abundant; judgment is becoming the moat. ## Where I focus Consumer industries — CPG, retail, travel, tourism, hospitality and airports. Within them: strategy and commercial transformation, marketing, sales, pricing and revenue growth management, and the AI-enabled operating models that tie them to the P&L. ## Background MBA, Faculty of Management Studies, University of Delhi. Two decades across Accenture, Strategy&, BCG, Genpact and Wipro. ## On stage I speak and host on AI and the future of consumer business, including at TEDx and Google events. To discuss a keynote or session, email [anubhav.pateriya@gmail.com](mailto:anubhav.pateriya@gmail.com) or connect on [LinkedIn](https://www.linkedin.com/in/anubhavpateriya?ref=anubhavpateriya.com). ![Anubhav Pateriya speaking at TEDx](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/06/anubhav-pateriya-tedx.jpg) ![Anubhav Pateriya in a fireside chat at a Google event](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/06/anubhav-pateriya-google-think-fireside.jpg) ![Anubhav Pateriya delivering a keynote on stage](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/06/anubhav-pateriya-keynote.jpg) ![Anubhav Pateriya presenting to a panel of speakers](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/06/anubhav-pateriya-panel.jpg) ![Anubhav Pateriya on a conference stage](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/06/anubhav-pateriya-conference.jpg) ## A note on views The views on this site and my social channels are my own, not Wipro’s. Everything here draws on public reporting and twenty years of industry experience — never on client engagements or confidential information. Companies are discussed solely on the basis of public sources. Nothing here is professional, legal, or investment advice. ### Speaking URL: https://www.anubhavpateriya.com/speaking/ Last updated: 2026-08-18T04:20:12.000Z **AI is solving the driving problem. The leaders who win are the ones who decide where the machine should go.** I help consumer leadership teams find that decision and act on it before their competitors do. ## Signature keynote ### The Destination Problem: Winning When the Consumer Enterprise Drives Itself AI is solving the driving problem. Demand is sensed rather than surveyed, supply chains rebalance themselves, and agents are starting to do the shopping. But a self-driving enterprise does not choose its destination. Drawing on what Walmart, Kroger, PepsiCo, and Marriott are doing right now, Anubhav Pateriya shows why execution is going abundant, why judgment is becoming the moat, and how the leaders pulling away decide where to point the machine before they switch it on. The audience leaves with a framework for the three destination choices only leadership can make. **What the room walks away with** - The three destination questions: a framework to separate the choices AI can make from the ones only leaders can. - A sharper read on where margin and moat move as execution commoditizes. - Current, concrete examples across CPG, retail, and travel they can act on immediately. - The one diagnostic question to put on their own board agenda. ## Also speaks on - **When the Shopper is an Algorithm:** winning the agent's pick as commerce goes agentic (retail). - **Decision Engineering, Pricing Power in the AI Era:** where the moat lives when everyone rents the same optimizer (CPG and revenue growth management). ## Built for Consumer C-suite and boards across CPG, retail, travel and hospitality; leadership offsites and strategy summits; and the industry stage, including NRF, Shoptalk, Groceryshop, the Consumer Goods Forum, and Skift. ## Seen on stage TEDx · Google keynotes · Beam Suntory · Business Today and Accenture · the BCG and Google report launch. ## About Anubhav Anubhav Pateriya is Managing Partner and Global Head of Consumer Consulting at Wipro, leading strategy and AI-driven transformation across CPG, retail, travel and hospitality. He has spent twenty years on every floor of the consumer enterprise: strategist at BCG, builder at Accenture, and operator running a nine-digit P&L. He writes on the Destination Problem and the self-driving consumer enterprise. [Get in touch](mailto:anubhav.pateriya@gmail.com?subject=Speaking%20enquiry) ### Privacy Policy URL: https://www.anubhavpateriya.com/privacy/ Last updated: 2026-07-10T23:52:37.000Z *Effective July 10, 2026* This site is operated by Anubhav Pateriya. It exists to publish essays, and it collects as little about you as possible. **What is collected.** If you subscribe, your email address — nothing else. The site also uses the standard functional cookies of its publishing platform (Ghost) to manage sign-in and subscriptions. There are no advertising trackers and no third-party analytics that identify you. **How it is used.** Your email is used for one thing: sending you new essays when they publish. It is never sold, rented, or shared with anyone. **Where it lives.** The site runs on Ghost(Pro), hosted by the Ghost Foundation, which processes subscriber data on my behalf. See [ghost.org/privacy](https://ghost.org/privacy/?ref=anubhavpateriya.com) for their practices. **Your choices.** Every email includes an unsubscribe link, which takes effect immediately. To access or permanently delete your data, email [anubhav.pateriya@gmail.com](mailto:anubhav.pateriya@gmail.com) and it will be done promptly. **Changes.** Any updates to this policy will be posted on this page with a new effective date. ## Posts ### Your AI agents are waiting for delegation of authority URL: https://www.anubhavpateriya.com/ai-agents-delegation-of-authority/ Last updated: 2026-08-18T01:14:42.000Z ## Somebody chose Balanced, and nobody wrote it down Somewhere in your company, in the last year or so, a meeting happened that nobody minuted. A vendor walked a team through the setup screens for a pricing agent, and about forty minutes in, on a screen with three choices on it, somebody chose Balanced. The setting decides how much evidence the agent waits for before it moves a price, and how far it is allowed to move it. Not Conservative, which felt timid, and not Aggressive, which felt like something you would have to explain later. Balanced. The room moved on, because there were eleven more screens. The person at the keyboard was probably a pricing analyst, and they were doing their job. Nothing about that afternoon was careless. In the same building, on the same day, that analyst could not have signed a fifteen thousand dollar contract without two names above theirs on the paperwork. Put those two afternoons side by side. The contract needed two signatures because fifteen thousand dollars is real money. The setting needed none, and it governs how that agent behaves across a portfolio for a year. On a large consumer business the pot behind it runs into the billions, which makes a single point of movement worth tens of millions, and I will show that arithmetic in a moment. So it reads as a story about weak controls. It is not. The controls are working as designed. What that afternoon settled is how much authority the machine has, and what it may do before it has to ask. Every decision it takes from then on happens inside that boundary. ## You bought agents to move faster, then set them to walk Start with what the money was for. Nobody approved an agent program in order to have a machine ask permission more politely than the last one, in fluent and tireless English, around the clock. The case was decision throughput, and the gap is worth stating without a multiplier attached: an agent can evaluate every item in every account every day, while a commercial team reviews a category once a month and gets to the accounts that shout loudest before the ones that matter most. That gap is what you bought. Almost nobody is capturing that gap, and the reason is not the technology. In the absence of a written mandate, nobody has the standing to widen anything, because a limit nobody approved is a limit you will personally own when it goes wrong. So whatever was chosen on the day stays chosen, the exception queue grows into a second full-time job for the person the agent was bought to free up, and the organization pays enterprise prices for a system left at walking pace. > Autonomy you have paid for and never authorized is a subscription. So let me put a number on it, and be clear about what kind of number it is. Nobody has studied what agent limits are worth, because the practice is barely eighteen months old. Nothing that follows is a finding. It is an estimate from the programs I have run, and you should discount it the way you would discount anyone else's, including the ones that arrive with a decimal point and a footnote pointing at nothing. The test is never who is saying it. It is whether they will tell you where it came from. Here is where mine came from. Across twenty years of commercial programs in large consumer businesses, what I have seen when trade allocation genuinely improves, not the plan but the allocation, is two to four points of spend moving out of promotions that lose money and into promotions that pay. Below about five billion in revenue I have seen a good deal more, because the starting point is usually worse and less of the budget is locked into habit. On a twenty billion dollar business, trade spend in the high teens of gross sales puts the pot near four billion, so a point of mix is worth about forty million dollars. Run it on your own numbers, and if your post-promotion analysis says something different, use yours. Agents do not raise that ceiling. They change how often you can act on it, which is worth nothing until somebody writes down how far they may go. So the question is not whether your controls are strong enough. It is whether they are specific enough to let you go faster on purpose. ## Your approval rules are not broken. They are watching transactions, not settings. Every company of any size owns a delegation of authority schedule, the grid that sets out which role may commit the company to what, up to which number, and who has to counter-sign. It was drafted by a general counsel and a finance director in a week neither of them enjoyed, and it gets opened about twice a year, once when a new director joins the board and once when something has already gone wrong. You will hear it said that agents will bleed a company by a thousand cuts, each one too small for anyone to notice. It is a good line and, if your schedule is in decent condition, it is mostly wrong. Each role almost always carries two limits, one for a single transaction and one for the month or the year, for exactly the reason you would guess: somebody worked out long ago that four modest invoices attract less attention than one large one. The thousand-cuts problem was solved decades ago, by people who never met an agent and were guarding against something far older. The agents are also better behaved than their reputation. In pricing this year the working pattern is that the agent proposes and a person disposes. On a Monday morning a category manager opens a queue of recommendations, works through the exceptions the system has flagged, accepts the rest, and anything touching a major customer goes up to somebody with a name and a bad week if it turns out badly. Every step carries an audit trail. You might object that these systems are hardly unwatched, and you would be right. They get reviewed constantly, in performance decks, vendor QBRs and exception reports. Problem is who sits in those reviews. The people reading the numbers sit in technical and compliance functions, and nobody in that room has the commercial authority to change the setting. The review is technical but the decision is commercial, so whatever was chosen on the day survives every review of it. So the money is guarded and the steps are logged. What is not guarded is the thing that decided what would be in that queue in the first place. ## The real decision was taken weeks earlier, on the setup screen An analyst who clicks approve on eight hundred price recommendations has approved them. Nobody would say they decided them. A batch takes about as long as a single one and feels a good deal more useful, and that is not laziness, it is arithmetic, because nobody reads eight hundred of anything on a Tuesday. The decision that mattered was taken back on the setup screen. What the agent may touch, which customers sit inside its scope, how much evidence it needs before it moves, what it protects when two goals pull against each other. Those four answers are the agent's operating limits, and once they are set, every approval afterwards happens inside the boundary they drew. The person clicking approve is choosing between options that somebody else already narrowed, at a meeting with no minutes. And that earlier choice appears nowhere on the schedule. There is a row for capital spend, a row for contracts, a row for hiring, a row for writing off bad debt, and no row at all for who may set the operating limits of a system that will touch several thousand prices before Christmas. ![Timeline showing the setup screen in week zero marked as the only point a commitment was made, followed by the agent running all year, weekly batch approvals, and a year-end result where the budget lands on plan but the mix is wrong.](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/08/fig_authority-timeline_in-body.png) Approvals continued all year. The decision was made once, in week zero. Nobody decided to leave that row off the schedule. You could object that none of this is new, and you would be half right. The last generation of enterprise software encoded commercial policy too, in credit limits, pricing conditions and approval workflows, all set by implementation teams in rooms much like that one. What has changed is what sits downstream of the setting. In an ERP a person still executed each transaction, so a badly chosen rule met a human being before it met a customer, and somebody usually noticed. A pricing agent removes that step. The setting is the decision now, several thousand times over, with nobody in between. I have spent a good part of two decades on the consulting side of large consumer and retail businesses, and I read these schedules early in an engagement, usually in the first fortnight, because they tell you more about how a company works than the strategy deck does. I have never seen one with a line for this. I would be glad to be sent one. That would matter less if the work sitting inside those limits were trivial. ## Pricing, promotion and markdown agents are already running inside those settings The work is neither trivial nor hypothetical. It is unglamorous, which is exactly what makes it easy to miss, and it runs at a volume no team could match. On trade promotion, an agent watches an event from the first day instead of the post-mortem, catches an underperforming mechanic while the money is still in the room, and adjusts inside the parameters the customer contract already allows. PepsiCo has published what this looks like at full scale, in a peer-reviewed account of two systems it runs called PromoAI and PricingAI. One searches millions of combinations of product, promotion and timing to build the promotional calendar5. The other sets base prices across the portfolio from estimated price elasticities. That is the trade calendar and the price list, computed rather than argued. On pricing, it weighs a competitor move, the stock position, the days left in the season and the margin target at the same moment, instead of in four spreadsheets owned by three people, one of whom is on holiday. In service it clears the small credit in ninety seconds instead of ninety hours. On the retailer's side of the same desk the picture is a mirror image. This month Dunelm, the UK homewares chain, put a shopping agent into its iOS app, built on Google Cloud's Gemini, letting customers describe a room in their own words and get a tailored set of products back6. Ask what decides which products come back and you are asking a merchandising question, not a technology one. The same shift is running through markdown timing and personalized offers, which means the negotiation between a brand and its largest customer is picking up two automated participants, each pushing toward a different goal at a speed neither commercial team can follow, while the joint business plan that governs the relationship still gets renegotiated once a year over lunch. Every one of those decisions is defensible on its own, and nothing in any single action would fail a review. The question is what all of them together do to the largest discretionary number on the P&L. ## Budgets check the total. Nothing checks the mix. Take trade spend, because it is the clearest case. Let us be precise about the word mix first, since it does the work here: mix is not how much you spend, it is where the money lands, meaning which accounts, which packs, which mechanics, at which depth. Trade spend is the largest pot of discretionary money a consumer goods company controls, and it has an odd property. It never appears as its own line in the published accounts, because it comes off gross sales before any other cost is counted. Industry estimates put it in the high teens as a share of gross sales, and since nobody is required to disclose it, those estimates come from advisers rather than filings. Now the part every revenue growth manager already knows. Nielsen has found for years that more than seventy percent of trade promotions do not break even4. Agents did not create that, and I am not going to pretend they did. It survives because every control around trade spend was built to answer one question: did we spend more than we said we would. The agent makes that old problem faster and much harder to see forming, because it optimizes against what it can measure, and some things are far easier to measure than others. A deep price promotion in a high-velocity account produces a clean, attributable lift inside a week. Investment in a slower account, or in a pack that builds a category over three years, shows almost nothing in the same window. Point an agent at measurable return and it will drift toward the first and away from the second, every week, without ever breaching a parameter, and without once mentioning that it is doing so, because nobody asked it to. So picture the year. The budget lands within a point of plan, every event sat inside approved parameters, every price move was inside the band, the audit trail is immaculate, and the business still ends up with more volume, thinner margin and two strategic accounts quietly underinvested. The review pack will be entirely green. Cumulative limits catch overspending and audit trails prove each step was authorized by somebody entitled to authorize it. Not one of those controls was built to ask whether nine hundred correct decisions add up to the mix you meant to hold, because until recently a human sat in the middle whose whole value was noticing when they did not. > None of it was settled in a strategy session. It was settled on a screen with three options on it. That noticing is what you are now buying back, and the only place to install it is in the settings themselves. Which is why the setup screen is not an IT decision. It is where your commercial strategy is now written, in a vendor's user interface, by whoever happened to be free that afternoon. You might reasonably ask whether somebody is going to solve this on your behalf. This month the industry answered. ## The new agent standard leaves permissions to you, deliberately On August 6, 2026, OpenAI, Amazon, Microsoft, Vercel and Cursor published Agent Plugins 1.0, a shared standard1 for how an agent's skills and tool connections get packaged so that any system can read them. Google joined as a core maintainer the same day. Five companies who agree on almost nothing agreed on this in about the time it takes a large enterprise to schedule a kickoff. What the standard covers is the packaging. What it leaves out, by its own stated scope, is who may install one of these things, what it is allowed to reach once installed, and what stops it doing damage. It sets out how an agent carries its tools. It says nothing about what the agent may do with them. That gap is deliberate, and the people who wrote it were right to leave it. Nobody outside your building can set your permissions, for the same reason nobody outside your building can write your trade terms, and anybody offering to would be handing you a template with somebody else's appetite for risk already loaded into it, usually somebody with no customers of their own to lose. What an agent may do to a price, a promotion or a customer relationship is a statement about what your company believes, and a standards body has no business having a view on that. So take it as the notice it is. The plumbing arrived faster than anyone expected. The permission did not arrive, and it is unlikely to. I argued in [The Judgment Layer](https://www.anubhavpateriya.com/the-judgment-layer/) that this part has no vendor. Here is the specification agreeing, in writing. If no standard is going to set your number, the next instinct is to find somebody who has set theirs and copy it. That turns out to be harder than it sounds, and the one company that has published a number is more interesting for what it did afterwards. The clearest published number comes from an adjacent industry, and where it comes from turns out to be part of the point. Airbnb said this month that its AI assistant now resolves close to forty-five percent of inquiries that start with it without a human agent, up from forty percent the previous quarter, and that support cost per booking fell about sixteen percent year on year2. Most people read that as a cost story. Read the two numbers together instead. Forty, then forty-five, one quarter apart. Airbnb has not said it widened a limit, but the assistant runs in more than fifty languages now with voice next, so somebody decided it could handle more and the number followed. The loop is still the part worth building, because it answers the objection that stops most of these conversations before they start. Nobody can tell you the right limit for your business on day one, and waiting until you are certain means waiting forever. > You do not need the right number. You need a defensible starting number, a review date, and an agreed rule for what would let it widen. Notice who did the disclosing, though. A platform business whose investors expect operating detail every quarter. It is hard to picture a manufacturer or a grocer telling the market what share of its pricing ran without a person, and harder still to picture an investor relations team volunteering it. So the benchmark you would like to copy probably does not exist, and there may be no industry average to point at when your board asks whether your number is sensible. Which makes the real question not what the number should be, but who is entitled to set it, and who is coming back to move it. ## Decide who is allowed to choose Balanced, then go about widening it There is a reason to settle this before the next deployment rather than after it. Gartner expects more than forty percent of agentic AI projects to be cancelled by the end of 2027, and the causes it gives are escalating costs, unclear business value and inadequate risk controls3. Not one of those is a model problem. These programs are not dying because the agent could not do the work. They are dying because nobody would put a name to what it was allowed to do while doing it, so the work never widened, never produced a number worth defending, and drifted until a finance review put it out of its misery. Which makes writing those limits the opposite of an administrative fix, and the wrong thing to hand to whoever manages the vendor. How much of your commercial position a machine may set is a delegation of authority, and it belongs where every other change to that document goes: approved by the board or the executive committee, recorded the same way, reviewed on the same cycle. I have argued before that judgment has to be a governed artifact rather than a settings screen. This is where that starts: one workflow, one page, one name. A company does not get an enterprise judgment layer by declaring one. It gets there by writing enough of these that they have to agree with each other. Three things make it real, and the third is the one companies skip. I will keep the examples in revenue growth management, where the money is most visible, though none of it is particular to that function. **Write the limits, including how they widen.** Take the workflow that is live rather than in pilot and put its limits on one page. What it may touch, by category and by customer. How much of the portfolio it may move in a period. How certain it must be before it commits, which is a statement about your appetite for being wrong and not a technical setting. What it escalates, to whom by name, within how many hours. What the executive committee sees monthly, including how many decisions it made and what share a person overrode, because the override rate is the one number here a pilot cannot flatter. And the line most companies leave out: what would earn it more room, stated in advance. If overrides stay low for two quarters and the outcomes hold, the limit moves, by a stated amount, on a stated date. A limit with no widening rule never gets widened, because nobody wants to be the person who proposed loosening a control that has not yet failed. ![Four-step loop for a revenue growth management agent: set a limit, watch overrides, check outcomes, widen it, returning to the start on the review date. A marker between checking outcomes and widening reads: most decision makers stop here.](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/08/fig_widening-loop_in-body.png) The widening loop, as a worked example for a revenue growth management (RGM) agent. **Record them where authority lives.** Put that page on the schedule as its own row, with an approval level beside it and a review date. Not a function. A person, the way that document has handled every other commitment for a century. **Then make the organization believe it.** A row changes nothing on its own, because the next implementation is already booked and the consultant running it will ask a room to pick a setting, exactly as last time, and somebody in that room will feel the pull of the middle option all over again. Nobody who configures these systems has ever needed sign-off for a configuration, and nothing in their training says this one is different. It has to be said out loud, more than once, to the commercial teams, to procurement and to the vendors, until the room knows that the person best placed to answer is not in it. If you cannot fill one of those lines, you have not found a technology gap. You have found a commercial position your company has never taken, and the agents will take it for you this quarter, in whichever direction the vendor thought reasonable. Writing that page is the cheapest thing you will ever do to make an expensive system earn out. No platform, no program, no new function. Your company has known how to put a name on a commitment for a hundred years. It has never once thought of a setting on a screen as one. Somewhere in your building, somebody chose Balanced. They may well have been right for that quarter. The questions are whether anyone decided they were the one to choose, and whether anybody is coming back to widen it. *Read next:* [*The Judgment Layer*](https://www.anubhavpateriya.com/the-judgment-layer/) *·* [*The Destination Problem*](https://www.anubhavpateriya.com/the-destination-problem/) *·* [*The line you won't cross*](https://www.anubhavpateriya.com/the-line-you-wont-cross/) *·* [*Decision Engineering*](https://www.anubhavpateriya.com/decision-engineering/) ## Sources 1. Agent Plugins 1.0, published 6 August 2026 by OpenAI, Amazon, Microsoft, Vercel and Cursor, with Google as a core maintainer. The specification covers packaging only and explicitly excludes distribution, permissions, provenance and sandboxing. [source](https://thenextweb.com/news/openai-agent-plugins-open-standard-skills-mcp?ref=anubhavpateriya.com) 2. Airbnb second quarter 2026 results and management commentary on AI-assisted customer service. [source](https://news.airbnb.com/airbnb-q2-2026-financial-results/?ref=anubhavpateriya.com) 3. Gartner, on agentic AI project cancellations by the end of 2027, citing escalating costs, unclear business value and inadequate risk controls. [source](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027?ref=anubhavpateriya.com) 4. Nielsen's analysis of more than one million UPCs and 39 million promotional events, covering $555 billion of US retail sales across 150 retail banners, found that almost three-quarters of promotions do not break even. Other industry findings put the figure nearer two-thirds. [source](https://www.foodnavigator.com/Article/2014/10/09/Three-quarters-of-CPG-promotions-don-t-break-even-Nielsen/?ref=anubhavpateriya.com) 5. PepsiCo's PromoAI and PricingAI, published in the INFORMS Journal on Applied Analytics, 11 June 2026\. PromoAI pairs machine-learning promotional forecasts with mixed-integer linear programming to build promotional calendars across trade channels; PricingAI sets base prices using Bayesian hierarchical models of own and cross-price elasticities. Both are optimization systems operating at scale rather than conversational agents. [source](https://pubsonline.informs.org/doi/abs/10.1287/inte.2025.0302?ref=anubhavpateriya.com) 6. Dunelm's Ask Dunelm shopping agent, live in its iOS app on Google Cloud's Gemini Enterprise. [source](https://www.retailgazette.co.uk/blog/2026/08/dunelm-launches-ai-shopping-assistant-with-google-cloud/?ref=anubhavpateriya.com) ### The Judgment Layer URL: https://www.anubhavpateriya.com/the-judgment-layer/ Last updated: 2026-08-12T19:43:17.000Z ## Accuracy was never the job For the better part of two decades, forecast accuracy was the industry's holy grail, and not without reason. It sits at the top of almost every planning scorecard, and even a ten to fifteen percent improvement measurably lowers inventory and lifts service. So the field chased it the way you chase anything that genuinely pays. Fortunes went into it. Careers were built on shaving points off the error, and a whole software market grew up selling the next decimal place. The catch is one you would have heard from any number of seasoned planners, usually moments before being sent off to improve the forecast anyway. Forecast accuracy measures how well you predicted demand. It says nothing about how well you handled the trade-offs that predicting it was supposed to serve: service against cost, inventory against risk, this customer against that one. A company can forecast beautifully and still decide badly. Accuracy was always necessary, and it was never the actual job. Accuracy is also harder to win than the software brochures suggest, and for reasons no model can remove. New products have no history to learn from. Promotions distort the very signal they create. A rival's stockout lands in your data as a surge of loyal new demand that evaporates the moment their shelf is refilled. And the bullwhip effect, the way a small wobble in store sales swings into a violent lurch at the factory, a modest uptick becoming a panic order becoming idle production lines a quarter later, guarantees that some of the noise is structural rather than solvable. The discipline even keeps score on itself, through a measure called Forecast Value Added that asks the deflating question of whether all that elaborate forecasting actually beats a naive guess. The answers are not always kind. I should say where I am standing. I have spent a good part of two decades on the consulting side of this, in the war rooms of large consumer and retail businesses, and I have run the accuracy programs myself. The ones that paid were the ones where we also changed what happened after the number arrived: who was allowed to act on it, how quickly, and without asking whom. Where we only improved the forecast, the improvement largely stayed in the forecast. That is the point at which I stopped treating accuracy as the deliverable. I once helped bring a new whiskey to market, where the volume you commit to is decided years before the first bottle is sold and no amount of modelling shortens the wait. The question in that room was never what the forecast said. It was how wrong we were willing to be, and in which direction. None of this means forecasting failed. It means forecasting matured, and mature things deliver diminishing returns. The next point of accuracy now costs what the first ten did and returns rather less. AI has genuinely pushed the frontier out, reading weather and web traffic and the general mood of the market alongside last year's sales. But somewhere in the last few years the binding constraint quietly moved. The question stopped being how well we can see what is coming. It became what we do about it, and how fast. ## The money leaks between knowing and doing Here is the part that should interest anyone who still believes the value lives inside the forecast. It mostly does not. It lives in the distance between the forecast and the decision. Incisiv's 2026 research on supply chain responsiveness gave that distance a name and a price: the latency tax, the margin lost between the moment a demand signal changes and the moment the organization actually acts on it. Their figure lands north of five cents on every dollar, which for a ten billion dollar business is five hundred and fifty million a year, currently filed under nobody's job. Treat that figure with the caution any number deserves when it is produced somewhere near a vendor. Halve it, then halve it again, and what remains is still the largest line item in the business that nobody owns. The precise size matters less than the absence of an owner, which is the part that should interest a board. Look at where the loss actually sits. Not in the not-knowing. The forecast was fine. The loss is that it arrived on a Tuesday, waited for the planning meeting, then for an approval, then for the next cycle, and by the time anyone moved the inventory, rebooked the freight, or shifted the promotion, the moment had already left the building. A perfect forecast acted on too late is not intelligence. It is an expensive historical document. That gap is exactly what this wave of technology is built to close. ## Show, know, do: three verbs, all rentable Before we look at what closes that gap, it is worth noticing what this industry has actually been selling. Step back, and thirty years of it resolve into a tidy story about a single verb. First we bought the software that could *show*. The great age of the dashboard, when visibility was the promise and a screen full of your own operation felt like control. It was real progress, and a quiet lesson in the limits of visibility. A dashboard shows you exactly what is happening and asks nothing of you in return; you can watch a problem develop in real time, in high resolution, and still be left to decide, entirely on your own, whether to do anything about it. Seeing was never the same as deciding. Most large companies built a control tower somewhere in the last decade, usually with a wall of screens and a ribbon to cut. Ask what decision came out of it last Tuesday and the room goes quiet in a very particular way. Then we bought the software that could *know*. Forecasting and demand sensing got genuinely good, and the promise moved from seeing the present to anticipating the future. Now we are buying the software that can *do*. This is the agentic wave, and it deserves a plain definition, because the word is working overtime this year. Ordinary AI predicts and then waits for a human to click. An agent is the part that does not wait: give it a goal and some guardrails and it reads the signal, weighs the trade-off, and acts across your systems, reordering, reallocating, rerouting, while a person supervises rather than steers. Gartner expects that by 2031, sixty percent of supply chain disruptions will be resolved with no human in the loop at all. Show, know, do. Each verb was, in its moment, sold as the decisive edge. Each verb, in its moment, became something any competitor could rent from the same short list of vendors by the same Friday. ## Autonomy arrives for everyone at once For a few years, speed did look like the whole answer. Close the gap, move before the other side does, win. This is where that conclusion, that the winners will simply be the most autonomous, quietly falls apart. An advantage that arrives for everyone at once is not an advantage. It is a utility bill. When your rival can sense, decide, and act in seconds on the same rented models you use, speed no longer separates anyone. The whole industry is about to become fast and autonomous together, which makes fast and autonomous the cost of staying in the game rather than the thing that wins it. *Do,* like *show* and *know* before it, is quietly becoming plumbing. So follow the logic to its end. We have commoditized seeing. We have commoditized knowing. We are commoditizing doing. When the machinery of sight, prediction, and action can all be leased by anyone with a budget, what is actually left for one supply chain to be better at than another? ## What is left is judgment What is left is the judgment about how all that rented horsepower gets used. Gartner calls the destination decision-centric planning, which is a polite way of saying the plan was never the point. The decisions it feeds are. I would go one step further. The scarce asset is [the judgment those decisions encode](https://www.anubhavpateriya.com/decision-engineering/), and it hides inside a single, deceptively small setting that sits on every agent: how sure it has to be before it acts. Anyone who drives a Tesla has felt this setting in miniature. Its self-driving software lets you choose a temperament. In Chill mode the car waits for a wide, unmistakable gap before it changes lanes; it would rather sit patiently behind a slow truck than commit to a move it is not fully certain about. Switch it to Mad Max and it slides into openings Chill would have waited to confirm, passing on thinner evidence and changing lanes far more readily. Nobody at Tesla had to call it Mad Max. Somebody chose that, and I think about it more often than I should. Same car, same sensors, same road. The only thing that changed is how much certainty it demands before it commits. That is not a driving preference. It is a judgment about the cost of being wrong weighed against the cost of being late. Now put that same dial inside your supply chain. A product spikes on a Tuesday. Real shift in demand, or a viral blip that is gone by Thursday? Turn your agents up to Mad Max and they commit inventory on day one, sometimes catching the entire wave, sometimes yanking stock across three regions to chase a ghost, at machine speed and full scale, before anyone has read the alert. Somewhere there is a warehouse holding four pallets of something that was briefly famous on the internet, still waiting for its moment. Leave them in Chill and they wait for the signal to confirm, which spares you the ghost but hands the wave to a faster rival and pays the latency tax in full. There is no factory setting that is simply correct. The right amount of doubt to tolerate before committing is different for strawberries than for washing machines, different in a supply shock than in a calm quarter, different for the flagship account you will protect at any cost than for the one that quietly absorbs the shortfall while everyone is asleep. Those are not settings a vendor can ship with sensible defaults. They are the encoded temperament of one specific company: its appetite for being wrong, the lines it will not cross, the trade-offs it has genuinely decided rather than merely inherited. Give a fleet of fast agents a clear judgment layer and they behave like one company with a point of view. Give them a vague one and you get a thousand confident reflexes firing in slightly different directions, each impeccably logged, all with your name on the outcome. ## Judgment has to be governed, not configured The first time I helped a company write its judgment down, it was not an AI project. It was an outsourcing one. When a large consumer goods business moves its planning to a service provider, somebody has to convert two decades of instinct into a document: what the team may decide on its own, what it must escalate, which customer gets served first when there is not enough to go round. It sounds like paperwork. It is not. It is the first time anyone has been made to say out loud what the company actually believes. The exercise is reliably uncomfortable. A good deal of what gets written down turns out to have been living in the heads of three people who have been there long enough to have opinions, and close enough to retirement that nobody has thought to ask them for the file. Those documents were crude, and they were built for an entirely different purpose. They were also, in a real sense, the first judgment layers: a written, owned statement of how this company decides, sitting outside the people who happened to be in the room that year. What agents change is the stakes. The old document governed a few hundred human decisions a week, with a supervisor on hand to catch the occasional absurdity. The new one governs several thousand machine decisions an hour, with nobody in the loop at all. Which tells you what a judgment layer actually has to be, and it is not a settings screen. It has to be a governed artifact. Someone owns it by name. It carries a version history, so that when the appetite for being wrong was widened in March, it is possible to see that it was, and who did it, and what they were worried about at the time. It is reviewed on a cadence, because the right posture in a tariff shock is not the right posture in a calm quarter, and a company that discovers this during the shock has already paid for the lesson. And it is auditable afterwards, so that when ten thousand agents did exactly what they were told and the result was ugly, the conversation is about the instruction rather than the machine. That is a better conversation, and a considerably shorter one. None of that is a technical requirement. All of it is a governance one, which is why this ends up on the board's agenda rather than the CIO's. You might reasonably ask why this layer does not commoditize like the other three, and in one sense it will: every vendor will soon sell a place to record your judgment. None of them can sell you the judgment, which is the difference between renting the safe and owning what goes inside it. This is the real product thesis of the decade, and it is a departure. The platform that wins from here is not a forecasting engine with better math, nor an automation engine with more connectors. It is a judgment system: the place where a leader's operating philosophy is written down once, precisely enough that ten thousand agents can execute it consistently, everywhere, without a meeting. The forecast is rented. The agents are rented. The judgment is the company. ## Where I would put the money It is easy to describe a shift and never take a position on it, so let me say where I would put the money. My bet is that this category is won on governance rather than on models. The engines will keep improving, and the improvements will keep arriving for everyone at roughly the same time, which is precisely why they will keep failing to separate anyone. The durable product is the one that lets a company state what it believes, install it across thousands of agents, revise it deliberately when conditions change, and show afterwards exactly what it believed at the moment a decision was made. Built properly, that is not a layer sitting on top of a planning system. It is the system, and the planning is what sits underneath. A bet is only a bet if it costs something. This one says the next point of forecast accuracy is worth less than the ability to explain a decision six months after it was made. That is not a popular sentence to say out loud in a room full of very good data scientists, and I have said it in a few. I would also want to be wrong in one specific way. If most companies turn out to be perfectly content on vendor defaults, and those defaults are good enough that the gap between two competitors never reaches the numbers, then judgment is a smaller prize than I think and scale wins instead. I do not believe that. It is still the first thing I would test. ![The AI-era supply chain: a gold Judgment Layer governs a swarm of identical rentable AI agents (the autonomous engine). The judgment defines the winner.](https://storage.ghost.io/c/5d/c9/5dc96621-b880-4c3b-8598-5623c532588f/content/images/2026/07/fig_judgment_FINAL_site_nofooter.png) ## The top job gets promoted, not replaced If that is where the value goes, the leadership job moves with it, and in a direction worth welcoming. For a generation, running a supply chain meant being the fastest, most tireless human in the loop, the person working exceptions at midnight so the plan could survive contact with reality. The machine is taking that job, and doing it without complaint, without fatigue, and without ever needing the weekend. What the machine cannot take is the decision that sits above the job: how much risk this enterprise will carry to move early, and which promises it treats as sacred when everything is on fire at the same time. Someone still has to author that, deliberately, in plain language, on the record. It is now the highest-leverage thing a leader does, and in most companies it is still running on whatever default a decade of quiet hesitation left behind. The genuinely good news is that the interesting part of the work is the part that survives. The machinery will handle the seeing, the knowing, and increasingly the doing. What it hands back, sharpened and impossible to duck, is judgment: [choosing where this enterprise is trying to go](https://www.anubhavpateriya.com/the-destination-problem/), and drawing the lines it will not cross to get there. [Execution is becoming abundant](https://www.anubhavpateriya.com/self-driving-never-self-deciding/). Judgment is the scarce thing, and it always was. The machines have simply been considerate enough to clear away everything that used to hide it. So here is the question worth carrying into the week. When the agents in your business finally start moving on their own, what is the judgment they will be executing, and did anyone actually decide it? For years, a slow and elaborate planning process was also a comfortable place to avoid answering that. There was always another meeting to call, another cycle to wait for. The agent era quietly removes the meeting, and with it the hiding place. Judgment is about to become very hard to fake. ### Self-driving, never self-deciding URL: https://www.anubhavpateriya.com/self-driving-never-self-deciding/ Last updated: 2026-08-12T19:43:17.000Z I recently bought a Tesla Model Y, and somewhere in the middle of ordinary traffic this two-ton robot made the future of AI agents crystal clear in a way no slide ever has. I have spent a while now telling consumer leaders that their enterprises are quietly going self-driving, that execution is about to become abundant and judgment the scarce thing that separates winners. I believe every word of it. But there is a difference between believing an idea and watching it happen at speed, in live traffic, with you in the passenger seat. The car threaded a genuinely chaotic stretch of road, an unprotected left, a cyclist drifting out of the bike lane, a delivery van double-parked precisely where it had no business being, and it did all of it with a composure I could not have managed on my third espresso. If a two-ton machine can read a live city street on its own, then whether software agents can run a supply chain has quietly stopped being a question of if and become a question of when. But the detail that stayed with me was not the driving. It was a small menu on the screen. ## Full Self-Driving is a subscription. The judgment isn't. Here is a thing worth sitting with. FSD is sold as a subscription, which means I am, in the most literal sense, renting an agent by the month. And yet the rented agent is not where my choices live. The car lets me decide how it drives. There is a gentle setting called Chill. There is Standard, the sensible middle. And there is one, I promise I am not inventing this, called Mad Max, which changes lanes with the enthusiasm of a man sprinting for a flight he was never going to catch. Same car, same self-driving brain, completely different conduct on the road, and the difference is not the technology. The difference is the instruction I hand it before I pull out of the driveway. That distinction is the whole story of where business is going. The car handles the execution immaculately. What it cannot do, and never will, is decide how bold I feel like being, or where I am actually trying to end up. I type the destination. I pick the mode. The machine drives; I judge. And if I select Mad Max and then grumble about the ride, that is not a defect in the software. That is a me problem. Now take that arrangement and set it down inside the company you run. ## $234 billion says the screen is disappearing On July 1, Gartner published a forecast with a number in it that made a good many software executives put down their coffee: up to 234 billion dollars of enterprise software spending is "at risk" between now and 2030, which by then is roughly one dollar in every five that companies spend on business software. The phrase they use is "agentic arbitrage," which sounds like something a hedge fund does at three in the morning, but the idea underneath is plain. Agents are beginning to do work straight across your systems, with no human clicking through the screens in between. The planner no longer opens the tool, squints at a grid, and nudges a few cells; an agent reads the signals, makes the call, and updates three systems downstream in the time it used to take to find the right tab. For twenty years we bought software you log into, trained people to live inside it, and priced the whole industry on how many humans sat in front of how many screens. When the agent does the work, those screens do not vanish so much as sink into the background, becoming plumbing that keeps the water running and that nobody thinks to open. Call it the disappearing dashboard, arriving at the office. ## The real disruption skips the IT department It is tempting to file this under software budgets and move on, which would be an expensive misread, because the action is nowhere near the tech team. Look at what is already running, not on a keynote slide but on real loading docks. PepsiCo now operates a supply chain with more than 1,500 AI agents alongside its human planners, and one of them watches the inbound dock, notices when pallets are piling up faster than they clear, calculates how badly that will choke the outbound flow, locates spare labor, and reroutes the crew, on its own, while the planner who once did that by walking the floor now supervises the agent that does. Walmart says it has spent years building the same muscle across retail, logistics, and software. General Mills went further and [attached a figure to it](https://www.anubhavpateriya.com/the-number-came-first/), three billion dollars of savings by 2030, with AI named out loud as a lever meant to deliver it. Notice what these companies are actually paying for, underneath the announcements. Not the software, because everyone rents the same agents from the same short list of vendors, and renting the identical engine as your rival has never once been an edge. What they are buying is the [judgment they hand the agents](https://www.anubhavpateriya.com/decision-engineering/): what to prioritize, what to protect at any cost, which trade-off wins when two goals collide at two in the morning and no one is awake to break the tie. That judgment is the Chill-or-Mad-Max dial for the whole enterprise. The tools are commoditizing. The instructions are becoming the company. ## Cluttered screens were great at hiding bad calls Here is the part that is faintly uncomfortable and, once you see it, quite funny. For years a busy screen was a superb hiding place for a decision nobody had actually made. As long as the promotion calendar looked complicated enough, no one paused to ask who decided that we always defend this account's margin and never that one's, because the sheer clutter did the work of camouflage, and strategy could stay comfortably vague so long as the dashboard looked sophisticated. Agents peel that camouflage off. An agent does not absorb your vague intent and split the difference the way a tired team does at quarter-end; it does exactly what you told it, at full speed, everywhere at once, which means a fuzzy instruction becomes fuzzy at industrial scale, in real time, in production, with your name on it. So a fair number of leaders are about to learn something slightly awkward: that half of what they always filed under "system settings" were never settings at all. They were strategy decisions wearing a disguise, and no one made them on purpose; they accrued, one reasonable click at a time, over a decade. The agent era hauls each of them into daylight and asks, with great politeness, "Is this genuinely what you want? Because I am about to do it ten thousand times before lunch." ## The machinery drives. It won't pick the address. I have taken to calling this the [self-driving consumer enterprise](https://www.anubhavpateriya.com/the-destination-problem/), lowercase, because it is a direction of travel and not a product you can buy. The machinery increasingly drives itself: the forecast runs, the price moves, the truck reroutes, the promotion fires, and the execution that used to swallow most of a leader's week turns abundant and very nearly free. But a self-driving car, however astonishing, is only ever as good as the address you give it. It will get you there beautifully. It will not decide, on your behalf, whether the trip was worth taking. Which is why the 234 billion dollar number is really a leadership story in a software costume. When the screens fall quiet and the agents take the wheel, the last human job in the building turns out to be the highest one: choosing where this enterprise is trying to go, and drawing the lines it must not cross to get there, however inviting the shortcut looks on the map. Execution is becoming a commodity. The destination is not. ## Before you hand over the keys You do not need a task force for this. You need one honest afternoon. Take a single workflow you are handing to agents this year, pricing or replenishment or trade spend, whichever is furthest along, and ask a question that has nothing to do with technology and everything to do with nerve: if this agent did exactly what we told it, ten thousand times, with no human quietly softening the edges, would we be proud of where it took us? If the honest answer is "not sure," you have not found a software problem. You have found a strategy you never quite wrote down. The good news, and I mean this, is that the top job has not been this interesting in years. The machine is finally quick enough to take the driving off your plate, which for the first time leaves you room for the part only you can do. So here is the easy one to sit with this week, the question that fits on a sticky note: what is the one thing your business should never do, no matter how good the numbers on it look? Answer that in a sentence, and you have already chosen your mode. --- **Read next:** [The Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/) · [Decision Engineering](https://www.anubhavpateriya.com/decision-engineering/) · [The number came first](https://www.anubhavpateriya.com/the-number-came-first/) ### The AI found the hotel. Someone else got paid. URL: https://www.anubhavpateriya.com/the-ai-found-the-hotel/ Last updated: 2026-08-12T19:43:00.000Z Two numbers came out of Adobe's analytics this summer, and they point in opposite directions. On Prime Day, shoppers who arrived at retail sites from AI assistants were 40% more likely to buy than shoppers arriving from search, email, or social. The machine out-converted every human channel. In travel, measured the same way by the same firm, visitors referred by AI assistants convert 28% less than average. Same technology. Same year. Same yardstick. Opposite results. The gap between those two numbers is one of the most useful things a consumer leader can study right now, because it shows exactly where the money moves when AI starts doing the shopping. ## A booking, step by step Watch what actually happens in travel. A traveler asks ChatGPT for a quiet hotel near Lisbon's old town, good for a couple, under 300 euros a night. The assistant suggests a boutique property. Fifteen rooms, family run, exactly right. Now comes the part that decides who gets paid. The traveler does not book. They open Google and type in the hotel's name. Adobe found 62% of consumers do exactly this, a quick double-check, before spending money on an AI recommendation. The search returns the hotel's own website, and above it a Booking.com listing with the familiar blue branding, four thousand reviews, and free cancellation. The stay gets booked on Booking.com. The hotel pays the platform a commission, typically 15 to 25 percent of the room revenue. Here is the uncomfortable part. That hotel did everything the industry is telling it to do. It rebuilt its website so AI systems could read it: structured room descriptions, machine-readable rates and availability. That work is the reason the assistant found it and recommended it. The hotel funded the discovery. An intermediary collected the sale. ## Trust starts decisions. It does not finish them. Adobe's travel data explains why this keeps happening. Visitors who arrive at travel sites from an AI assistant stay 70% longer and leave 41% less often than average visitors. They show up nearly decided, because the assistant already did the comparing. And they still convert 28% less. The explanation is not technology. It is trust. People trust AI enough to start a decision, not enough to finish one. A phone charger, fine. A two thousand dollar anniversary trip, no. So between the recommendation and the payment, travelers insert the double-check. And that double-check runs through the places they already trust with their money: Google, the big booking platforms, review sites. That is the first thing worth writing down. In any category where the purchase feels risky, the AI recommendation does not end the customer journey. It starts a second step where the buyer looks for reassurance. Whoever owns that step, the moment trust finally completes, is who gets paid. One more detail matters. In late 2024 that travel conversion gap was 86%. It is now 28%. Consumers are learning to trust machine recommendations, fast. The open question is who that maturing trust will belong to when it arrives. ## The incumbents are not confused. They designed this. The two companies that dominate online travel bookings are Booking Holdings and Expedia Group. They are the largest online travel agencies in the world, the middlemen who sell other companies' hotel rooms and flights and keep a cut of each sale. Both have told investors exactly how they see the AI shift playing out. On an April earnings call, Booking Holdings CEO Glenn Fogel said the company is working with OpenAI, Google and other AI providers so that Booking's inventory shows up wherever a traveler starts their journey. In early May, Expedia CEO Ariane Gorin went further. AI, she said, can be a powerful discovery layer, but the actual booking is best handled by a trusted, scaled provider. She meant Expedia. She added that if AI discovery evolves into a paid advertising space, that is a game Expedia knows well. Then Expedia did something worth pausing on. At its partner conference in May, it announced that it will let any other company plug its own AI travel assistant directly into Expedia's rooms, flights, and payment system. Read that carefully. Whoever builds the travel assistant of the future, Expedia is offering to be the invisible checkout inside it. Both Booking and Expedia are already bookable through ChatGPT. None of this is a reaction to AI. It is a position, taken early, by companies that have seen this movie before. Which brings us to the history. ## We have seen this movie In the 2010s, hotels learned an expensive lesson about Google. Showing up on page one was not a strategy, because the space above the search results was sold to the highest bidder, and the highest bidder was usually an online travel agency spending billions on marketing. A guest would search for a hotel by name and still end up booking through a middleman. The hotel created the demand. The platform taxed the transaction. Twenty years of that spending built the online travel agencies something durable: brand recognition, deep review libraries, content everywhere on the web. And here is the quiet advantage, that asset base transferred straight into the AI era without any new investment. AI models recommend what they have read the most about. The companies with two decades of accumulated web presence score highest on exactly the signals the models reward. The distribution war of the 2010s is not being disrupted by AI. It is being re-run, one layer up, at machine speed, with the same teams already holding the strong positions. ## Why retail flipped and travel did not So why did the same AI traffic become the best-converting channel in retail and the worst in travel? Because of where the sale actually closes. In retail, the assistant hands the shopper to a store's checkout, the purchase is small, and the sale completes right where the shopper lands. The store that got discovered is the store that gets paid. In travel, the purchase is expensive and emotional, the double-check kicks in, and the sale migrates to whichever platform holds the most consumer trust. The business that got discovered watches the sale complete somewhere else. That is the rule hiding inside Adobe's two numbers. The higher the stakes of your product, the more the AI era rewards whoever owns trust at the moment of payment, not visibility at the moment of discovery. ## This is not a travel story Every consumer business is currently being told to get AI ready. Clean product data, structured attributes, content a machine can parse. Do that work. A product an AI cannot read does not exist in that channel. That is the argument I made in [The Agent Channel](https://www.anubhavpateriya.com/the-agent-channel/): the terms of the new shelf are set by whoever the machine buys through. But being readable is not a strategy. It makes you visible inside a funnel, and the real question is who owns the bottom of it. A packaged-goods brand optimizing its product content for AI assistants should ask where the resulting purchase actually closes: on a marketplace that keeps the margin and the customer data, or somewhere the brand controls. An airline watching assistants compare fares should ask who completes the ticket. The hotel industry is simply the first to see the full shape of the trap, because it has walked into a version of it once before. ## The destination decision AI has solved the finding problem. It can locate the product, compare the price, and walk the customer to the edge of the purchase. What it has not decided, what it cannot decide, is where the sale is allowed to close and on whose terms. That is the [Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/), and in most consumer companies it is currently being made by default, while the teams work through the AI-readiness checklist. The two largest middlemen in travel have already told the market, plainly and in public, which end of the sale they intend to own. The discovery race is real and everyone has to run it. But the race pays out at the close. That is where the next set of winners is being decided, right now. --- **Read next:** [The Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/) (the cornerstone thesis) · [The Agent Channel](https://www.anubhavpateriya.com/the-agent-channel/) · [Both sides of shopping just went commodity](https://www.anubhavpateriya.com/both-sides-of-shopping-just-went-commodity/) ### The first Prime Day the machine out-converted every human channel URL: https://www.anubhavpateriya.com/the-agent-channel/ Last updated: 2026-08-12T19:43:18.000Z For thirty years, the referral channel that sent a retailer its best customers was a settled question: search, then email, then social, roughly in that order. Prime Day 2026 broke the order. For the first time, the highest-converting traffic a retailer received did not come from a person clicking a link. It came from a machine that had already done the shopping. ## The number is not the story The headline figure was a record. US shoppers spent $26.4 billion online across all retailers over the four days of the event, June 23 to 26, up 9.3% year over year, according to Adobe's final tally. Summer is starting to rival the holiday season. But the record is the least interesting thing that happened. The interesting thing is who was doing the buying, and how they arrived. ## Twelve months, from the worst channel to the best Shoppers who landed on a retail site from an AI assistant were 40% more likely to make a purchase than shoppers arriving from paid search, email, or social media. That is Adobe measuring visits that turn into orders. A year ago, the same traffic converted 23% worse than those channels. In twelve months, the assistant went from the weakest-converting referral source a retailer had to the strongest. Not incrementally better. A reversal. The mechanism is not mysterious. A person who arrives from a search ad is still deciding. A person who arrives from an assistant has already been walked through the comparison, the price check, and the shortlist. The agent did the deliberation before the click. What lands on the site is intent that is nearly closed. ## Why a sub-1% channel decides the next decade Here is the objection, and it is a fair one. AI traffic is still a low single-digit share of visits at most retailers. It is a rounding error next to search. That is precisely why this is the moment that matters. Channel economics get written before channels get big. The take rates, the data-sharing terms, the ranking rules, and the ad loads of a channel are all set while it is still small and everyone is experimenting. By the time a channel is large, those terms are load-bearing and nearly impossible to renegotiate. The retailers deciding their posture toward the agent channel now, while it is under one percent, are the ones who will set the terms rather than inherit them. And the two largest retailers in America have already placed opposite bets. ## Two of the biggest retailers, opposite bets Walmart and Target opened their catalogs to outside assistants. Walmart wired its inventory into ChatGPT and into Google's shopping surfaces; Target built a conversational experience inside ChatGPT; both adopted the open protocol that lets an assistant read a merchant's full product feed. The logic is simple. If the machine is doing the shopping, the job is to be readable, priced, and available wherever the shopper's agent happens to look. Meet the agent on the agent's ground. Amazon went the other way. It blocked ChatGPT's crawlers from reading its listings. It sued Perplexity's shopping agent for logging into Amazon accounts on a shopper's behalf and won a preliminary injunction, now under appeal at the Ninth Circuit. And it pointed the more than 250 million users of its own shopping assistant back to Amazon.com. The logic here is just as sound, and it is financial. Advertising is now an estimated third of Amazon's operating income, and advertising depends on shoppers browsing Amazon, seeing sponsored placements, and choosing. A shopper who delegates the decision to an outside agent is a shopper Amazon no longer monetizes on the way to checkout. ## The tell And then, as the event opened, Amazon bought ads inside ChatGPT to promote Prime Day. Sit with that. The company fighting hardest to keep the agent out of its own aisles paid to show up inside the agent everyone else is using. It is not a contradiction. It is the whole strategy in one move: keep the transaction on your turf, but do not cede the discovery surface where the next generation of shoppers is forming intent. Even the fiercest opponent of the agent channel is already renting space in it. ## This is not a retail problem It is tempting to file this under "Amazon versus everyone" and move on. That would be a mistake, because the same shift is arriving in every consumer category, not just retail. A CPG brand's product page is about to be read by an agent before a human ever sees it, scored on price, availability, ratings, and substitution rules, and either recommended or quietly left off the shortlist. A hotel's rate and room availability will be parsed by a booking agent that never sees the lobby photography the brand spent a fortune on. An airline's fare will be compared inside an assistant that has no loyalty to the carrier. In each case the choice is the same one the retailers just made: be natively legible in the AI layer, or be resold inside it and hand over both the margin and the customer relationship. ## The driving problem is solved. The destination problem is not. This is what I mean when I say AI is solving the [driving problem](https://www.anubhavpateriya.com/the-destination-problem/) and leaving leaders the harder one. On both of the paths above, the driving is done. The AI can find the product, compare the price, and complete the purchase. The capability is not in dispute and it is not a differentiator, because it is available to everyone. What the machine cannot decide is whose terms the transaction happens on. Open your shelf to every agent, and you gain reach but rent your growth from platforms you do not control. Keep the shopper direct, and you own the relationship but you are betting your assistant knows the customer better than the ones they already talk to every day. Neither answer is wrong. But it is an answer only a leadership team can give, and it gets compiled straight into your commercial terms, your data policy, and your media mix, the same way [strategy now gets encoded](https://www.anubhavpateriya.com/decision-engineering/) into the pricing room. That is not a technology decision. It is a destination decision, and this Prime Day was the first time it showed up in conversion data instead of a strategy deck. ## The question on your desk By next Prime Day, the agent channel will be a line in your trade math, whether or not you chose to put it there. The retailers and brands that picked their answer this year will be negotiating from a position. The ones who did not will be inheriting someone else's terms. So the question is not whether your customers will shop through an agent. Adobe just showed they already convert better when they do. The question is whether you have decided, on purpose, whose terms that sale happens on. --- **Keep reading** [The Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/) is the cornerstone: AI solves the driving problem; leaders own the destination. This essay extends two threads from it: [Both sides of shopping just went commodity](https://www.anubhavpateriya.com/both-sides-of-shopping-just-went-commodity/) and [Decision Engineering: AI's first win is in the pricing room](https://www.anubhavpateriya.com/decision-engineering/). ### The line you won't cross: pricing's new moat in the AI era URL: https://www.anubhavpateriya.com/the-line-you-wont-cross/ Last updated: 2026-08-12T19:43:18.000Z **For twenty years, revenue growth management chased one dream: price every shopper individually, at exactly what they will pay. Price-pack architecture, promotion optimization, elasticity models. All of it was a workaround for the thing we could not do at scale.** AI just delivered it. And that is the trap. ## The reckoning arrived in 2026 Look at what 2026 has already produced. In January, California's attorney general opened a sweep into businesses using personal data to set individualized prices. This spring, Congress opened an investigation into surveillance pricing in travel. New York's Algorithmic Pricing Disclosure Act is now in force. More than twenty states have bills moving. The FTC has named algorithmic pricing a 2026 enforcement priority. And when lawmakers warned that AI fares could climb to each traveler's personal "pain point" and named Delta, the airline disputed pricing on individual data. The warning landed anyway. ## When a capability becomes free Here is what actually changed. AI did not make pricing faster. It made individualized, personal-data pricing trivial and invisible. The capability that was scarce for two decades is now close to free. And when a capability becomes free, it stops being the advantage. What is left is a choice: whether, and where, to use it. That choice is now the whole game. ## The line is the decision Point the optimizer at maximum extraction and it will price to the pain point. It will move the quarter. It will also spend down what the model never sees: customer trust, and the benefit of the doubt from a regulator with twenty other cases open. Adoption without a point of view does not just make you average anymore. It makes you a target. So the pricing decision that matters this year is not how much more you can take. It is what you forbid the engine to do. Which segments you protect. Which data you refuse to price on. Where you draw the line yourself, on purpose, before someone draws it for you. That restraint used to look like leaving money on the table. In 2026 it is the opposite. It is the cheapest brand insurance you can buy, and it is a moat, because the competitor running the same optimizer at full extraction is one headline away from losing both the trust and the case. Execution is becoming abundant. The machine can price anyone. Judgment is becoming the moat, and in pricing, judgment is the line. Where is the line you will not let your pricing AI cross? And did you draw it on purpose, or are you waiting for a regulator to draw it for you? This is the Destination Problem in the pricing room: as execution goes abundant, judgment becomes the moat. The fuller thesis is [here](https://www.anubhavpateriya.com/the-destination-problem/), and the companion piece on pricing as the encoded decision is [Decision Engineering](https://www.anubhavpateriya.com/decision-engineering/). ### Decision Engineering: AI's first win in consumer goods is in the pricing room URL: https://www.anubhavpateriya.com/decision-engineering/ Last updated: 2026-08-12T19:43:19.000Z **AI's first durable win in consumer goods did not show up on the shelf. It is showing up in the least glamorous room in the building: the one where trade promotion and pricing get decided.** ## What PepsiCo actually did Look at what PepsiCo is doing. It is pushing AI into revenue growth management: optimizing trade spend, sharpening price-pack architecture, and using first-party data to target promotions and "next best action" at the store level. In April it even put a global VP in charge of artificial intelligence and what it calls "Decision Engineering". Sit with that phrase for a second, "Decision Engineering". A ninety-billion-dollar company just called out where the value is, and it is not the model. ## Why pricing is the real prize Here is why pricing unlock is real; trade promotion is one of the largest lines on a consumer-goods P&L and, for decades, one of the most wasted. The reason was never bad intent. It was math no human team could hold: every SKU, every pack, every account, every week, against every competing price. RGM ran as a quarterly committee debate over a handful of scenarios. AI makes it continuous and granular, and the number it moves is net revenue realization, not volume for its own sake. Now the part most decks end up missing. If everyone rents the same optimizer and points it at the same syndicated data, RGM commoditizes. Two competitors converge on the same prices and the same promotions and compete margin to the floor. Adoption alone is not an edge here. It is a faster route to average. And optimizers do exactly what you tell them. Aim one at short-term trade ROI and it will deliver short-term trade ROI, quietly spending down what it cannot see: brand equity, price tolerance, the loyalty that lets you hold a premium next year. ## The edge is the decision you engineer So the edge is not the engine. Everyone can rent the engine. The edge is the proprietary read on how your shoppers actually respond, and the tradeoff rules you engineer into the system: what you refuse to discount, where you defend premium, how you weigh margin against volume against loyalty on every line. That is the decision. That is the moat. I have spent twenty years in that room. The board question on pricing this year is not "are we using AI in RGM." Everyone will be using it. It is narrower and more uncomfortable: which tradeoffs are we encoding, and what will we never let the optimizer do? What is the one pricing decision in your business you would never hand to the machine? Answer that honestly, and you already understand Decision Engineering. Part of a larger thesis: [the Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/), on why, as execution goes abundant, judgment becomes the moat. First shared on [LinkedIn](https://www.linkedin.com/posts/anubhavpateriya%5Fais-first-durable-win-in-consumer-goods-activity-7476019351364530176-pe6W?ref=anubhavpateriya.com). Join the discussion there. ### The number came first URL: https://www.anubhavpateriya.com/the-number-came-first/ Last updated: 2026-08-12T19:43:02.000Z **Kroger, General Mills, and Coca-Cola each made a serious AI move this year. The three press releases read as unrelated. They are not.** ## Three different rooms of the P&L Kroger is pointing AI at e-commerce operating profit. It expects roughly $400 million of improvement in 2026 as it rolls out a Gemini assistant grounded in its own assortment, pricing, and availability data. General Mills has tied more than $20 million in supply-chain savings to AI demand sensing since 2024\. Coca-Cola is using AI in revenue growth management to serve premium and value shoppers at a precision no human team can hold. Three different rooms of the P&L: the storefront, the supply chain, the price. One move underneath, and it is not "they bought AI." ## The number came first Each one decided which number had to move before building anything. Kroger chose e-commerce operating profit. General Mills chose cost-to-serve. Coca-Cola chose price realization. The number came first. The technology came second. That ordering defined the destination and then execution. Now the contrast. The surveys keep landing on the same split: roughly three in four leaders call AI a top priority, but barely a third can point to a financial return, and more than half report no benefit yet. That is not a technology gap. Starting a pilot requires no decision. Naming the P&L line it answers to forces one, and most organizations quietly decline. ## The encoded decision is the moat Here is the part that should change how you fund this. None of the three is winning on the model. They rent the same engines their competitors can. What they own is the decision encoded into the system: which number it serves, on whose economics, using which proprietary data. That encoded decision is the moat, and it is the one line no competitor can buy. So the board question this year is not "are we doing AI." Everyone is doing AI. It is narrower and more uncomfortable: which decision are we encoding, in which room, against which number. AI is taking over the how. These three decided the where. That is the whole contest now. Part of a larger thesis: [the Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/), on why, as execution goes abundant, judgment becomes the moat. First shared on [LinkedIn](https://www.linkedin.com/posts/anubhavpateriya%5Fkroger-general-mills-and-coca-cola-each-activity-7474945686455734273-pUkJ?ref=anubhavpateriya.com). Join the discussion there. ### Both sides of shopping just went commodity URL: https://www.anubhavpateriya.com/both-sides-of-shopping-just-went-commodity/ Last updated: 2026-08-12T19:43:20.000Z **Over the past few weeks, two headlines crossed most consumer leaders' desks as unrelated. One was about how people shop. One about how AI gets procured. Together they retire an assumption the industry has run on for 20 years.** ## Demand: the journey is compressing to a single turn Start with behavior. The shopping journey is compressing to a single turn. Rezolve's data (one vendor, but consistent with the broader zero-click shift) is stark: seven in ten purchases now follow a single search, up from six in ten a year ago. The shopper stopped behaving like a database query and started behaving like a brief. ## Supply: the agent is becoming a rental Now the supply side. Serving that shopper used to take years and a platform team. Now it ships as a configuration. Amazon packaged the engine behind its $12B Alexa business into an AWS kit any retailer can wire to its catalog, citing 3.5x the conversion of keyword search. It is not alone: Google, OpenAI with Stripe, and Shopify are racing to turn the same capability into shared rails. The agent is becoming infrastructure, not advantage. ## The scarce resource flips Here is why they connect. When demand concentrates into one high-intent turn, the whole interaction lands in a single moment: what the agent recommends, on what logic, what it refuses to do. Just as that moment becomes everything, the machinery that runs it becomes a rental. The scarce resource flips. Advantage used to come from owning the funnel: traffic, tooling, optimization muscle. The funnel is now one click long, and the tooling is bought. What stays scarce is the judgment encoded into the agent. That breaks two things every consumer org budgets around: - **The metrics.** Click depth, dwell time, pages per session were proxies for a many-step journey. In a single-turn world they measure something that no longer happens; the dashboards and incentives built on them now optimize noise. - **The roadmap.** "Build our own assistant" assumed the build was the moat. When a competitor rents the same engine and goes live in weeks, the build is table stakes. The only durable choices left are configuration choices. ## The default-config trap Here is the trap, and it rhymes with every prior platform shift. The risk is not moving too slowly. It is moving fast and deploying the default. Two brands on the same engine, both out of the box, are interchangeable: same answers, same logic, same refusals. Adoption without a point of view does not differentiate you; it commoditizes you faster. What matters is the judgment the box cannot supply: which intent you optimize for, what you forbid the agent to do, how it weighs margin against loyalty. I call this the Destination Problem. AI is solving the driving problem. The choices that decide who wins are destination choices, and they belong to leaders, not the machine. Execution is going abundant. Judgment is the moat. The real question is not which agent you deploy. Everyone deploys, on the same rails. It is the destination question: when the engine is rented and the shopper arrives with a brief, what does your brand decide that the machine never can? Explore the full thesis: [the Destination Problem](https://www.anubhavpateriya.com/the-destination-problem/). First shared on [LinkedIn](https://www.linkedin.com/posts/anubhavpateriya%5Fover-the-past-few-weeks-two-headlines-crossed-activity-7473360681795207168-uRqB?ref=anubhavpateriya.com). Join the discussion there. ### The Destination Problem URL: https://www.anubhavpateriya.com/the-destination-problem/ Last updated: 2026-08-12T19:43:20.000Z For thirty years, advantage in consumer industries came from execution. Better forecasting, tighter supply chains, sharper campaigns, faster checkout. The winners were the ones who ran the machine best. With the onset of AI and agents, that machine is now starting to run itself, today in parts, tomorrow in full. And when execution becomes something you can buy, running it well stops being the advantage. The advantage moves to the one decision the machine cannot make: where to point it. I call this the Destination Problem. AI is solving the driving problem. The choices that decide who wins are destination choices, and they belong to leaders, not the machine. ## The consumer enterprise is becoming self-driving Look at what crossed consumer leaders' desks in just the past few months. Walmart is rebuilding its assistant into a system that plans the family's week of meals, fills the basket, and completes the purchase. Kroger has put a hard number on its AI storefront: roughly $400 million of e-commerce operating-profit improvement in 2026\. Amazon packaged the engine behind its multibillion-dollar Alexa shopping business into a kit any retailer can rent and stand up in weeks. Marriott and IHG opened their inventory to autonomous booking agents, and Booking.com's agent crossed 50 million monthly users. General Mills has banked tens of millions from AI orchestrating its supply chain, not from chatbots. Read together, these are not isolated technology announcements. They are the same shift, arriving from every direction at once: demand sensed rather than surveyed, supply chains that rebalance themselves, agents transacting with other agents, and increasingly, agents doing the shopping. This is the largest redesign of how consumer companies run since ERP. The consumer enterprise is becoming self-driving. ## A self-driving enterprise does not choose its destination Here is the tension every leadership team is about to feel. A self-driving car still needs a person to say where to go. A self-driving enterprise is no different. As AI takes over how the business runs, advantage shifts to the choices only leadership can make. The machine gets faster at getting there. It never decides where there is. And the timing is what makes this urgent. The capabilities are commoditizing in real time. When Amazon will rent your competitor the same shopping engine it rents you, the engine stops being the moat. Execution is going abundant. Judgment is becoming the scarce resource, and the scarce resource is where the margin goes. ## The three destination questions Strip away the technology and every consumer leader faces three destination choices. Not one of them can be handed to the platform or the pilot team. ### 1\. Which demand do you own when the shopper is an algorithm? The journey is collapsing to a single turn. When most purchases follow one search and an agent builds the shortlist, you are no longer winning the shopper's eye. You are winning the agent's pick. Shelf psychology does not travel into that layer; machine-legible value does. The choice is which intent you optimize to own, and you encode it into your data, your ranking logic, and your retail-media bids, not your brand deck. I worked through one version of this in [Both sides of shopping just went commodity](https://www.anubhavpateriya.com/both-sides-of-shopping-just-went-commodity/). ### 2\. What does your brand mean to a machine, and to the human behind it? An agent reads your brand in milliseconds as price, availability, ratings, and substitution rules. Two brands running the same engine, both straight out of the box, are interchangeable: same answers, same logic, same refusals. The differentiator is what your brand refuses to do, the judgment the box cannot supply. Adoption without a point of view does not differentiate you. It commoditizes you faster. ### 3\. Where does your pricing power live when execution is instant? When everyone can point the same revenue-growth optimizer at the same syndicated data, prices and promotions converge and margin drains to the floor. The edge is the proprietary read on how your shoppers actually respond, and the tradeoff rules you engineer into the system: what you refuse to discount, where you defend premium, how you weigh margin against loyalty. PepsiCo has gone as far as naming the muscle. It calls it Decision Engineering, the subject of [Decision Engineering: AI's first win is in the pricing room](https://www.anubhavpateriya.com/decision-engineering/). ## Strategy is no longer in the deck. It is encoded. Here is the part most leaders miss. In a self-driving enterprise, strategy stops living in the slide and starts living in the system. It gets compiled into the pricing agents, the supply-chain policies, the demand algorithms. Walmart's choice of which agentic rail to join is strategy. Marriott's decision to expose its inventory through an API is strategy. The single number a company tells its AI to move is strategy. Choosing where to go and building the machine that gets there are becoming the same act. I traced this through three companies in [The number came first](https://www.anubhavpateriya.com/the-number-came-first/). This is why the destination cannot be delegated to a transformation office. It is a leadership decision that gets compiled into code, and once it is compiled, it runs at machine speed whether or not anyone chose it on purpose. ## Why most will get this wrong Analysts expect a large share of agentic AI projects to be abandoned over the next two years. The reflex is to read that as the technology failing. It is not. Starting a pilot requires no decision. Naming the destination it serves, the P&L line it answers to, and the choices it is allowed to make, forces one, and most organizations quietly decline. The projects that die rarely had a destination to begin with. So the trap is not moving too slowly. It is moving fast and deploying the default. Two competitors on the same engine, both running the out-of-the-box configuration, end up in the same place: average, faster. The leaders pulling away are not the ones with the most agents. They are the ones who decided where the machine should go before they switched it on. ## What this looks like in your industry The pattern is the same everywhere, but it lands differently by sector. In **consumer goods**, the first durable win is not on the shelf. It is in revenue-growth management, where AI can move net revenue realization at a granularity no human team could ever hold. The moat is the decision you encode, not the model you license. In **retail**, the contest is shifting from winning the shopper to winning the agent that shops for the shopper. Whoever makes their assortment, price, and availability legible to that agent owns the basket. Everyone else becomes a substitutable line item. In **travel and hospitality**, every category is about to re-fight the disintermediation war it thought it had lost, this time against an agent rather than an online travel agency. Make your brand natively legible in the AI layer, or be resold inside it and hand over both the margin and the customer relationship. ## The questions to put on the table If you lead a consumer business, the board question this year is not whether you are using AI. Everyone is. It is narrower and more uncomfortable: - Which demand are we choosing to own when the algorithm is the shopper? - What will we never let our agents do, in order to protect what the brand means? - Which line on our P&L is the AI actually moving, and did we choose that line on purpose? Answer those honestly and you have your destination. Leave them to the machine and the machine will choose one for you. ## Where I sit I have spent twenty years on every floor of this building. Strategist at BCG, builder at Accenture, operator running a nine-digit consumer P&L, and now leading Wipro's global consumer consulting business across CPG, retail, travel and hospitality. The pattern from inside these transformations is consistent. The winners are not the ones with the best models. They are the ones who decided where the machine should go. Execution is becoming abundant. Judgment is becoming the moat. The only question that still belongs entirely to you is the destination one. What is the destination question on your desk right now? Read next: [The number came first](https://www.anubhavpateriya.com/the-number-came-first/), [Both sides of shopping just went commodity](https://www.anubhavpateriya.com/both-sides-of-shopping-just-went-commodity/), and [Decision Engineering: AI's first win is in the pricing room](https://www.anubhavpateriya.com/decision-engineering/). First shared on [LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7472633793493905408/?ref=anubhavpateriya.com). Join the discussion there.