Returns and sizing: the two questions agents ask that catalogues rarely answer.
Sizing and returns are the two questions that decide whether an AI agent will risk a recommendation at all, and most catalogues answer neither with any real specificity. "True to size" and "easy returns" aren't answers, they're the absence of one.
6 min read
Sizing: the fact that determines whether the order is right the first time
Say a parent asks an agent to find school shoes for a nine-year-old with wide feet, needs them by the weekend, wants a straightforward return if they don't fit. Two retailers stock a near-identical pair. One catalogue entry says "true to size." The other says: this style comes up narrow, wide-fit buyers should go up half a size, and returns are free and processed within two days of the item arriving back. The second retailer gets the recommendation, because it answered both questions the request depended on.
"True to size" is a claim, not a fact an agent can act on. It tells the agent nothing about whether this specific style runs narrow, whether a known wide or narrow foot needs to size up or down, or whether the sizing is consistent across the retailer's own range. Without that, an agent recommending the item is guessing on the buyer's behalf, and a returned pair of shoes is exactly the outcome an agent exists to help avoid.
Returns: the fact that determines how much risk the agent is asking the buyer to accept
"Easy returns" says nothing about what happens: how many days, who pays for the courier, whether a refund lands before or after the returned item is received. An agent weighing two similar products treats a vague returns policy as a real cost, because it's asking the buyer to accept an unknown level of hassle if the guess is wrong. A specific, favourable policy is a genuine point in a product's favour, not just fine print.
The same two questions surface just as sharply outside clothing. A retailer selling monitors that says nothing beyond "easy returns," with no dead-pixel policy stated, is asking a buyer to accept a different, unstated risk: whether a single dead pixel counts as a fault at all. An agent recommending a big, expensive, hard-to-return item weighs that unstated risk more heavily than it would on a £15 t-shirt.
Two sentences do more than two paragraphs
A longer paragraph about quality and craftsmanship doesn't add a fact an agent can check. The specific fact does: this style runs narrow, wide-fit buyers should size up, returns are free within thirty days, refund on receipt not on request. Two sentences that hold up as fact carry more weight than two paragraphs that don't.
Where this shows up for a business
A retailer wins this recommendation on the fact, not the phrasing: the real sizing pattern and the real returns terms, pulled from its own systems rather than rewritten as a friendlier paragraph. That's what Selfe connects, so an agent working through a request gets the specific answer instead of the vague reassurance most catalogues default to.
Do we need to guarantee returns to be recommended?
No, but the actual terms need to be specific and checkable rather than vague. A clear, ungenerous policy beats a vague, generous-sounding one.
What if our sizing genuinely is standard, with no quirks?
Say so specifically: this range runs true to size, consistent across the collection. That's still more useful to an agent than the generic phrase, because it's a checkable claim rather than a stock one.
Five questions an agent checks before it'll recommend a product.
An AI agent shopping on someone's behalf doesn't browse a product page the way a person does. Before it will recommend anything, it needs five separate questions answered: does the exact variant exist, is it currently in stock, will it work for this specific case, what happens if it's wrong, and will it arrive in time. A page built to persuade a browsing human answers few of those directly.
What makes one listing easier for an agent to sell than another.
Given two nearly identical products, an AI agent recommends the one with fewer unresolved questions, not the one with better marketing copy. The gap between them is almost always one missing, narrow fact, not a missing feature.
What "buyable" needs beyond stock.
An AI agent has already checked that a product is in stock. Buyable asks more of it than that: is the price the one that'll actually be charged, and can this specific item reach this specific buyer at all. Miss either and "in stock" stops meaning anything useful.