Ottom

Learn

How do I know if AI will buy on my website?

You find out by sending an agent and watching where it stops. An AI agent reads your store as data rather than as a rendered page, so nothing you can see in a browser tells you whether one can find your products, understand your prices and terms, or reach your payment step. The failure is invisible from the inside, which is why guessing is the one approach that does not work.

You find out by sending an agent and watching where it stops. There is no reliable way to answer this by inspecting your own store, because an AI agent reads your store as data rather than as a rendered page, and the difference between those two things is invisible from inside a browser.

That is an unsatisfying answer, so it is worth being precise about why the obvious alternatives do not work.

Why looking at your store tells you nothing

When you open your own product page, your browser does an enormous amount of work on your behalf. It runs the scripts that calculate the price. It waits for the image carousel. It renders the returns policy you set in a graphic. It remembers you are signed in. None of that happens for an agent, which reads the text and the structured data and moves on.

What a shopper sees against what an agent reads The same product page. A person sees an image, a price rendered by script, and a policy set in a graphic. An agent reads only the text and structured data, so the price and the policy are missing from its view. WHAT A SHOPPER SEES product photo Trail Jacket £129.00 In stock, ships tomorrow Free returns for 30 days WHAT AN AGENT READS image: no alt text Trail Jacket price: not found availability: not stated returns: image only
The same product page, twice. Your store looks fine in a browser, which is why this is invisible to you.

So the store looks fine. It is fine, for people. The parts that fail an agent are the parts your browser was quietly compensating for, which means the better your site looks to you, the less your impression is worth as evidence.

The four places it can stop

An agent buying on someone’s behalf works through four stages, and "will it buy" is really four separate questions. Failing any one of them ends the purchase, and failing an early one means you are never assessed on the later ones.

The four stages of an agent purchase attempt A vertical sequence: discover, then evaluate, then compare, then buy. An agent can stall at any one of the four, and a store that fails an early stage is never measured on the later ones. 1 Discover Find your store and your catalogue at all 2 Evaluate Work out what you sell, for how much, and whether it is in stock 3 Compare Weigh you against an alternative 4 Buy Get a real product to the payment step
An agent can stall at any of the four stages. A store that fails discovery is never measured on checkout.
  1. Discover. Can it find your store and your catalogue at all? This is decided by your robots rules, your sitemap and whether you publish a product feed.
  2. Evaluate. Can it work out what you sell, for how much, whether it is in stock, and what your terms are? This needs text and structured data, not layout.
  3. Compare. Can it weigh you against an alternative? If your shipping cost only appears at checkout, it has to guess your real total, and a guess is easy to lose.
  4. Buy. Can it get a real product to the payment step? Variant pickers, popups, forced account creation and bot walls all end journeys here.

The expensive case is the last one. A store that fails at discovery was never in the running. A store that passes discovery and evaluation and then fails at checkout was chosen, and then could not complete. That is a sale an assistant had already sent you.

What you can check yourself, today

None of these prove a purchase would succeed, but each one rules something out, and they cost you an afternoon rather than a project. Work through them in order, because an early failure makes the later checks moot.

  • Read a product page with JavaScript disabled. If the price is missing, see how to add product schema for AI agents.
  • Check your robots.txt for a blanket disallow on agent user agents. The robots.txt guide covers which categories are which.
  • Check whether your bot protection challenges an identifiable, well behaved agent. This is the bluntest failure of all: see bot wall.
  • Open your checkout on a phone, in a narrow window, signed out, and try to reach the payment step without creating an account.
  • Look for your shipping cost and your returns terms as text rather than as an image.

For the longer version of this, structured so you can hand it to somebody else, use the agentic readiness checklist or work through an afternoon audit.

Why a signal check is not the same as an answer

There are free tools that read your storefront and score the signals: whether you have structured data, whether your feed exists, whether your robots rules permit an agent. They are useful and they are cheap, and they answer a narrower question than the one you asked.

A signal check can tell you that your product markup is present. It cannot tell you that your variant picker only responds to a hover, that your checkout fails in an in app browser, or that your popup fires on the cart page. Those are behaviours, and behaviour only shows up when something actually tries. That distinction is the subject of agent readiness score vs purchase attempt.

The honest answer

You know whether an AI agent will buy on your website when one has tried and you have watched where it got to. Everything short of that is inference, and the inference is unusually unreliable here because the evidence you have easiest access to, your own eyes on your own store, is the evidence most likely to mislead you.

The next question is usually where the purchase would even complete, which depends on the assistant. Can ChatGPT buy from your store covers the case where the answer lands back on your own checkout.

Questions

Can I just look at my own site and tell?

No, and this is the trap. Your storefront renders for you because your browser runs the scripts, loads the fonts, waits for the carousel and shows you the policy graphic. An agent gets none of that. The gap between what you see and what it reads is exactly the thing you cannot see.

Is there a quick check I can do in five minutes?

Open your best selling product page with JavaScript disabled and read it as plain text. If the price, the stock status, the variant options and the returns terms are not all there in words, an agent has the same problem. It is a crude test and it catches a surprising amount, but passing it is not the same as completing a purchase.

Does this matter if AI shoppers are still a small share of my traffic?

That is a judgement about volume, and it is yours to make. The point of measuring is that being invisible becomes a decision you made rather than one that was made for you. Bot rules, theme updates and checkout apps all change this without telling you.

Scan your store