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.
6 min read
Price drift is the quieter failure
A kettle shows as in stock in a retailer's feed. It's genuinely sitting in the warehouse. It's also the 230V European version, not the UK plug the same page sells everywhere else on the site, the price hasn't been updated since a sale ended three weeks ago, and the colour an agent's been asked to find sold out yesterday in everything but the feed's stale count. In stock, technically. Not buyable, for this request, right now.
The piece on variants, stock and delivery covers why a business needs all three checked before recommending anything. Buyable adds two more failure modes on top of a technically accurate in-stock flag: a price that's gone stale, and a variant that isn't really sellable to this buyer even though the listing says it is.
A sale ends and a feed doesn't update for days. A price goes up and the cached figure an agent has access to is still the old one. Either way, the number an agent would quote isn't the number a buyer would pay, and recommending on stale pricing erodes trust in the recommendation itself, not just the one purchase.
The same drift runs in reverse, too. A flash sale marks a price down for six hours and a feed that only refreshes overnight keeps quoting the old, higher figure for the rest of the day, at which point a retailer isn't losing a sale so much as quietly talking an agent out of one it was actively trying to make.
A listing can be right about the product and wrong about the buyer
A feed marking the kettle "in stock" globally when it's only sellable, or only wired, for one region creates the same gap from a different angle: technically accurate, practically useless to a buyer outside that region. Fixing it isn't a longer disclaimer on the page. It's the feed reflecting the real, current constraint, region by region, before an agent ever recommends it.
Where this shows up for a business
Selfe checks price and region against what's actually sellable to this buyer at the same moment it checks stock, so "in stock" never gets treated as the whole answer. Getting the warehouse count right was never the hard part. Keeping the price and the sellable region current alongside it is.
Isn't stock the hardest part to get right anyway?
Often it's the easiest, most systems track it accurately. Price drift and regional mismatches are the parts that go stale silently, underneath a stock count that's still technically correct.
How would we even notice a mismatch like this?
Usually only when a customer complains after the fact. A live check catches it before an agent recommends the product, not after someone's tried to buy it.
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.
Variants, stock and delivery: what an agent checks before it recommends.
Three facts gate whether an AI agent will recommend a product at all, regardless of anything else about it: does the specific variant exist, is it there right now, and will it arrive in time. Get any one wrong and the rest of the listing doesn't matter.
What a product feed needs that a product page doesn't.
A product page is written for a person looking at one item. A product feed is what an AI agent reads to compare many at once, and it needs a different kind of precision: structured, decision-relevant fields, not the same copy reflowed into a spreadsheet.