--- type: Article name: "What it means for an AI to know a place, not just describe it." id: "https://selfe.ai/insights/known-and-matched/what-it-means-for-an-ai-to-know-a-place-not-just-describe-it" url: "https://selfe.ai/insights/known-and-matched/what-it-means-for-an-ai-to-know-a-place-not-just-describe-it" publisher: "did:web:selfe.ai" description: "A product description tells an AI agent what a shop says about itself. Being known is a different, higher bar: whether an AI can trust what it's told enough to act on it." datePublished: "2026-09-14T15:38:25.120Z" author: The Selfe team --- # What it means for an AI to know a place, not just describe it. > A product description tells an AI agent what a shop says about itself. Being known is a different, higher bar: whether an AI can trust what it's told enough to act on it. ## In short - A description is a claim with no source. Being known means the claim has a source, a date, and someone accountable for it. - "We say so" was never really enough, it just wasn't tested until an AI started comparing claims across shops rather than reading one at a time. - A known shop can compete on the strength of its real claims rather than its name, because an AI can check them rather than just repeat them. Anyone can write a sentence that sounds right, which is exactly the problem: an AI reading "handmade in Britain" has no way to tell a true claim from a well-written invented one. Being known is the harder thing underneath the sentence, a claim with a source, a date, and someone accountable for it, and it's what lets an AI move from repeating a description to recommending based on it. A business can be entirely truthful and still lose to a less honest competitor whose claims simply carry a source, because an AI has to check, not trust. ## Why "we say so" was never really enough Say a small leather goods maker describes its bags as "handmade in Britain." That's a description, true or not, and an AI agent reading it has no way to tell the difference. Being known means something more specific: the claim has a source, a date, and someone willing to stand behind it, "handmade at our Bristol workshop, verified by a supplier audit in June," rather than a sentence that reads the same whether it's true or entirely invented. This is the difference between a shop that's been described and a shop that's been known. Description is cheap: anyone can write a sentence that sounds right. Being known is the harder, more valuable thing underneath it, and it's what lets an AI move from repeating a claim to recommending based on it. The internet made every shop its own publisher, describing its own products in whatever terms sound best. That was never a problem for a person browsing casually. It becomes a real problem the moment an AI is choosing what to recommend on someone's behalf, because an assertion with no source behind it is indistinguishable from one that's simply well-written. "Handmade in Britain" and a fabricated version of the same claim look identical on the page. Only a source, a date, and someone accountable for the claim tell them apart. ## What makes a shop known, rather than just described Three things, stacked: the claim itself, stated specifically rather than generally; where it came from, a supplier record, a certification, a site visit, not just the shop's own say-so; and when it was last true, since a fact that was accurate a year ago isn't automatically accurate today. A shop that has all three for its real, important claims is known. A shop that only has the first, however well it's written, is only described. ## What this looks like when it's missing Say an AI is comparing the leather goods maker against a larger competitor for a "handmade in Britain" gift request. The competitor's claim carries a named workshop and a certification number. The maker's claim is just the sentence, however true it is. An AI answerable for its recommendation has no way to prefer the unsourced claim over the sourced one, even when the unsourced claim happens to be the more honest of the two. Being right and being known are not the same thing, and only the second one is visible to something that has to check rather than trust. ## Why this pays off before trust or bookings even enter the picture A known shop doesn't need to argue harder for a sale. It needs an AI to be able to check its claims quickly and move on to comparing them against a competitor's, rather than treating every claim as equally unverifiable and defaulting to whichever shop happens to be more familiar. Being known properly is what lets a smaller, less recognised shop compete on the strength of its actual claims rather than its name. That's the layer Selfe exists to build: turning a shop's real claims into something sourced, dated and checkable, so being known stops depending on how convincingly something is written and starts depending on whether it's true. ## Doesn't this just mean adding more detail to product pages? Not more, more specific. A shorter claim with a source beats a longer one without. ## What counts as a source, for a small shop with no formal certifications? A site visit, a supplier invoice, a dated photo, even a specific named person who can vouch for the claim. The bar is "someone accountable," not "an official body." ## Act on this - [Check whether an agent can buy from you](https://selfe.ai/agentic-commerce/buyability-check) — the free readiness scan. - [Discover venues](https://selfe.ai/api/registry/discover) — `POST`, semantic browse across the registry. - [Match a bookable answer](https://selfe.ai/api/registry/match) — `POST` with dates and party size. - [Verify Selfe's identity](https://selfe.ai/.well-known/did.json) — `did:web:selfe.ai`. - [Agent card](https://selfe.ai/.well-known/agent-card.json) — how to connect over A2A or MCP.