6 August 2026 · Discovery

Merchant Discovery: how AI assistants find local businesses

When someone asks an assistant for "a good sushi place near Rembrandtplein that's open now", no ranked list of ten blue links is involved. One or two businesses get named. This is what decides which ones.

The shape of the question changed

Classic local search returned a page and let the user choose. An assistant answers instead: it resolves the intent, queries whatever structured sources it can reach, and commits to a recommendation. The merchant either appears in that answer or is invisible — there is no second page.

Ranking position five used to mean reduced traffic. In an assisted answer it means no traffic.

What the assistant actually resolves

1. Entity match. Is this business a single, unambiguous entity? Duplicate or conflicting listings across sources are the most common reason a merchant is skipped.
2. Constraint check. Open now, price band, cuisine, accessibility, distance. These come from structured attributes, not from prose on a website.
3. Evidence of quality. Review volume, recency, and — increasingly — whether the merchant responds to reviews at all.
4. Retrievability. Whether the data is reachable through an interface the assistant can call, rather than rendered for human eyes only.

Where merchants have leverage

The uncomfortable part for most marketing teams: the lever they own most completely — website copy — is the one that moves the outcome least. The levers that matter are data hygiene problems wearing a marketing hat.

What we are building

Obenan's Merchant Discovery exposes merchant data through an interface assistants can query directly, so a business is answerable rather than merely published. If you want to test it against your own locations, reach out via obenan.ai.