Why price alone doesn't win an agent's recommendation.

What gets weighed alongside price

Say two shops both sell the same style of leather wallet, one for twenty-eight pounds, the other for thirty-two. The cheaper one looks like the obvious winner. But the cheaper shop's returns policy is a vague "contact us within a reasonable time", its stock count hasn't been confirmed recently, and nothing on the page says where the leather comes from. The pricier shop states its returns terms exactly, confirms real stock, and names its supplier. An AI recommending on someone's behalf, answerable for getting it right, has a real reason to prefer the second one, and it isn't the price.

This runs against how price comparison used to work, where the lowest number, all else being roughly equal, won by default because a person doing the comparing had limited patience to check anything else. An AI has no such limit. It can check the returns terms, the stock claim, and the sourcing detail as easily as it reads the price, which means all of those things are now genuinely part of what gets compared, not just a tie-breaker nobody has time to reach.

Three things most shops have never thought of as competing with their price tag: whether the returns policy is a specific, checkable term rather than a vague promise, whether the stock claim is current rather than assumed, and whether any claim about the product itself, material, origin, size, has a source behind it rather than just being stated. A shop that's clear and current on all three can lose on price and still win the recommendation, because the AI is weighing certainty as much as cost.

A vague policy used to read as normal. Now it reads as risky.

Most returns policies are written vaguely, so a vague one has never stood out to a person browsing a single page in isolation. Nobody's ever compared two policies side by side in the moment of deciding, because comparing that closely took more patience than most shoppers have. An AI makes exactly that comparison on every recommendation it produces, and a vague policy that used to blend in with every other shop's vague policy now sits directly next to a specific one from a competitor, on the same screen, in the same decision. The shop with the vague policy hasn't changed anything about how it writes its terms. What changed is that its terms are now being read next to someone else's, every single time, by something that never gets tired of checking.

Why this changes what's worth investing in

For a shop competing against bigger players on price, this is genuinely good news: the fields most worth tightening up, the returns policy, the stock accuracy, the sourcing detail, cost nothing to fix compared to competing on price directly. Writing a specific, honest returns policy instead of a vague one is an afternoon's work. Undercutting a larger competitor on price indefinitely isn't a strategy most small shops can sustain, and it isn't the only lever available to them any more.

This is the Trusted step from "What 'agent-ready' means" applied to price: a claim with a source behind it outweighs one without, price included. Getting this right means making sure the facts that get weighed, not just the price, are specific, current and sourced, wherever an AI is checking them from. That's the layer Selfe exists to build: connecting a shop's real stock, terms and product facts to something an AI can check directly, so a fair price with clear, trustworthy terms behind it can beat a lower price with none. For how we think about evidence over assertion, see the Trusting job in plain language.

So a higher price can win?

Yes, when the terms and facts around it are clearer and more current than the cheaper option's. Certainty is part of what's being compared, not just the number.

What if our returns policy really is flexible, case by case?

State the actual process, even a flexible one, as specifically as possible. "Get in touch and we'll sort it" is vaguer than "get in touch within 30 days, we typically offer an exchange or refund."