AI Commerce WatchTracking the agents that shop

Sep 15, 2026 - AI Commerce Watch

When AI Agents Start Buying: What Agentic Checkout Means for Your Store

The next bot that checks out on your store may have a human's credit card and permission behind it. Agentic commerce - AI assistants executing purchases for users - is early, funded, and incompatible with how most storefronts distinguish friend from bot.

The collision course

Storefront defenses are built to stop automation. Agentic shoppers ARE automation - legitimate, revenue-carrying automation that looks, to a naive bot filter, exactly like the enemy. Merchants face a fork: block agent traffic and lose a growing sales channel, or allow automation and reopen the door to hostile bots wearing an agent's clothes.

What breaks first

Bot policy. 'Block automation' stops working as a default; the question becomes which automation, verified how. Attribution. When an agent researches and buys in one machine-speed session, the marketing journey compresses to a point - last-click will credit 'direct' or the agent's referer, neither of which describes how the sale was won. Fraud screening. Velocity and behavior rules tuned for humans will false-positive on agents, and rules relaxed for agents will be exploited by bots imitating them.

What to do now

Merchants do not need an agent strategy deck; they need three practical capabilities. Visibility: detect and separately report agent-class traffic rather than lumping it with human or hostile traffic. Identity: support emerging agent-authentication standards as they stabilize, so good agents can prove themselves. Terms: decide what agents may do on your storefront - browse, cart, checkout - and enforce at the edge, where policy can keep pace with machine speed.

The takeaway

The stores that win agentic commerce will be the ones that can tell three kinds of automation apart in real time: hostile bots, beneficial crawlers, and paying agents. That is a detection problem first and a strategy problem second.

The AI Size Chart Measured a Different Garment

November 7, 2026

AI-generated size charts look complete and measure the wrong thing: a flat-lay photo, a different season's cut, or a size run the brand never sold. Returns follow the chart, not the garment.

A shopper buys a medium based on the store's AI-generated size chart, which says the chest measures 22 inches. The medium that arrives measures 20.5 inches, because the chart was generated from a flat-lay photo of last season's cut, and the model filled in the numbers it expected instead of the numbers on the tape. She returns it, orders a large, and the large is too big. Two shipments, one return label, and a customer who now buys a size up everywhere, just in case.

Size charts are the highest-stakes content on a fashion storefront, because a wrong number costs a return, and returns cost more than the margin on the item. AI generation is tempting here precisely because measuring every size of every style by hand is slow, and that is exactly why the generated chart cannot be trusted without a human holding a tape measure.

Where the numbers come from, and why they drift

A generated chart usually starts from a photo or a product description, not from the garment. The model estimates measurements from what similar garments usually measure, which means it reproduces the industry average instead of the actual cut. A boxy fit, a cropped length, or a vanity-sized run all disappear into the average, and the chart looks authoritative while being wrong in exactly the places that matter.

The drift compounds across sizes. If the model gets the medium wrong by an inch, it usually extrapolates the small and large from the same wrong base, so the whole size run is off by a consistent amount. Shoppers who size up based on the chart then get a garment that is wrong in the other direction, and the return rate rises on both ends.

The return math

A size-related return costs the store the outbound shipping, the return label, the inspection and restock time, and often the sale itself, because a returned item in a popular size sells again but a returned item in a slow size sits. When the chart is the cause, the same defect repeats on every order of that style, so one bad chart can generate dozens of returns before anyone connects them.

The signal is in the return reasons. If returns for a style cluster on too small and too big at the same time, the chart is wrong, not the customers. If they cluster on one direction only, the garment runs small or large and the chart is at least honest about it. Most stores never split return reasons by style, so the pattern stays invisible.

What a trustworthy chart needs

Measure the garment, not the photo. A human with a tape measure, one sample per size, chest, length, and sleeve, recorded against the actual production run. Publish the measurement method alongside the numbers, flat or on-body, so a shopper can compare with a garment she already owns.

Then keep the chart out of the AI pipeline. Size data is a measurement, not content, and it should live in the product record where a merchandiser owns it, not in a generated page that nobody reviews. When the cut changes, the chart changes with it, on the same day, by the same person who approved the sample.

Related reading: Glossary, About, AI Chatbots That Invent Discount Codes.

Frequently asked questions

Why are AI-generated size charts unreliable?

Because they estimate from photos or descriptions instead of measuring the garment. The model reproduces the industry average, so a boxy fit, a cropped length, or a vanity-sized run disappears into the average and the numbers look authoritative while being wrong.

How do I know if a size chart is causing returns?

Split return reasons by style. If returns for one style cluster on too small and too big at the same time, the chart is wrong. If they cluster in one direction, the garment runs small or large and the chart is at least honest.

What should a trustworthy size chart include?

Measurements taken by a human with a tape measure, one sample per size, chest, length, and sleeve, plus the method used, flat or on-body, so a shopper can compare against a garment she owns.

Should size data live in the AI content pipeline?

No. Size data is a measurement, not content. It should live in the product record where a merchandiser owns it, and change on the same day the cut changes, by the person who approved the sample.

Frequently asked questions

What is agentic checkout?

A purchase completed by an AI agent acting for a shopper: the agent finds the product, applies the shoppers preferences and payment details, and checks out, with the human setting the task rather than clicking through each step.

What should a store check first about agent purchases?

How its products appear to agents: whether catalogue data, prices, stock, and terms are machine readable and current. Agents buy from the data they can read, and stale or hidden data means being skipped or misrepresented.

Who is responsible when an agent buys the wrong thing?

That depends on the stores terms and the agents mandate, which is why both need writing down before disputes. Stores that define how agent orders are treated avoid absorbing every ambiguity personally.

Will agent checkout replace normal checkout?

Not wholesale. It suits repeat, well specified purchases first. Stores should expect a growing share of orders arriving through agents alongside ordinary shoppers, and design product data for both.