--- type: Article name: The surfaces where agents choose and buy. id: "https://selfe.ai/insights/market-surfaces-and-standards/the-surfaces-where-agents-choose-and-buy" url: "https://selfe.ai/insights/market-surfaces-and-standards/the-surfaces-where-agents-choose-and-buy" publisher: "did:web:selfe.ai" description: "Customers increasingly begin in places that help them research, compare and progress a purchase. The durable task is to make a business understandable and actionable across those surfaces without rebuilding everything for each one." datePublished: 12 September 2026 author: The Selfe team --- # The surfaces where agents choose and buy. > Customers increasingly begin in places that help them research, compare and progress a purchase. The durable task is to make a business understandable and actionable across those surfaces without rebuilding everything for each one. ## In short - The customer need stays the same even when the starting surface changes. - Discovery, decision information and transaction progress should not depend on a single interface. - Standards are useful when they improve the hand-off between systems. If a marketplace listing says a violin is in stock from a feed that synced three days ago, and the shop's own till knows the last one sold this morning, an AI agent checking either surface gets a different answer to the identical question. Only one of the two is true right now. A small musical instrument shop sells through its own site, a marketplace listing, and increasingly gets checked directly by AI assistants answering a shopper's question about a beginner violin. Four different starting points, and the shop only has one real answer to give: what's in stock, at what price, under what terms, right now. ## The journey no longer has one front door A customer may begin with a search result, an AI assistant, a marketplace listing, or the shop's own site directly. The starting surface changes, violin, £300, beginner might get checked through any of the four, but the underlying job stays the same: understand what's being asked, compare it against credible options, and take the next safe step toward a sale. For the shop, the aim isn't chasing every new interface that might send a customer its way. It's keeping the underlying stock, price and terms clear enough that whichever surface is asking gets the same true answer. ## Separate the surface from the source of truth A marketplace listing, an assistant's answer, and the shop's own product page can all explain or collect a request, but none of them should become the only place the shop's real stock and price live. If the marketplace listing says "in stock" from a feed that synced three days ago, and the shop's own till knows the last violin sold this morning, an agent checking either one gets a different answer to the same question. Availability, price, policy and fulfilment need one clear owner underneath all four surfaces. That's what lets a new channel, a new AI assistant integration, a new marketplace, become a connection task rather than a full rebuild of how the shop answers a stock question. ## Standards earn their place at the hand-off Technical standards matter when they make it easier for an AI assistant checking the violin's availability to ask the shop's systems the right question and get a specific answer back, not a generic product description. They're not proof of readiness by themselves, a shop can adopt every standard going and still have stock data that's three days stale underneath it. Judge a standard against the actual customer route: does it reduce the shop having to enter the same stock and price information separately for every surface that asks? Does it preserve who's accountable for confirming the sale once an agent has checked the facts? ## Build for the next useful question The shop that gets this right can answer the next sensible question wherever it's asked, is this specific violin, this specific size, available today, whether that question arrives through a search result, a marketplace, or an AI agent acting on a parent's behalf. The shop's real stock and pricing need connecting to every surface that might ask, not maintained as four separate, slowly drifting versions of the same answer. That's the specific work Selfe does, so the focus stays on the sale itself rather than on any one platform or protocol. ## Do we need to be present on every AI platform separately? No. The same underlying connection to real stock and pricing can answer the question wherever it's asked, rather than needing a separate, manually maintained version for each surface. ## What if our marketplace listing and our own site show different stock numbers? That's the exact problem worth fixing first, since an agent checking either one has no way to know which is current. ## 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.