Restricted analysis, made public daily.
Declassified under standing order Edition No. 050 Saturday, August 22, 2026

Verified by Yelp.
Disabled by Default.

Yelp’s moderators check the trade license, and the platform posts the shield for human visitors. In certain categories, the business pays for the feature. In the company’s own developer documentation, the same verified record carries one more sentence: “Access is disabled by default.” And the AI Chat endpoint, documented for conversational, agent-style queries, lists rating, review snippets, hours, and photos in its published response schema. No licenses field appears.

The shield on a Yelp page is not decoration. It means a human moderator at the platform checked the business’s trade license and confirmed it stood in good standing on the date they looked. Yelp’s own Support Center describes the work in plain terms, and adds one more detail worth holding onto: the Verified License badge is “a paid feature available to businesses in certain categories and regions.” The contractor with the shield didn’t just earn the license from the authority that issued it. In many cases they also paid the platform to confirm it.

Now read the other document Yelp publishes about the same record. The company licenses its business data to developers. The Model Context Protocol section of its data page for AI agents describes bringing “real-time Yelp data into your AI workflows.” In the developer reference for the Business Details endpoint, the verified license appears as a fully structured field: license type, license number, issuing authority, expiration date, even the date Yelp verified it. Five properties, machine-readable, already built. The field’s published description then closes with a sentence that does all the work of this edition: “Access is disabled by default. To have access, please contact us.”

This is not a policy aimed at licensure. The same default-off gating covers a whole shelf of partner-tier fields: messaging status, Request a Phone Call, Request a Quote. That is precisely what makes it worth filing: the license, the one fact in the record an issuing authority stands behind, sits on the same access shelf as a phone-call button. And the AI Chat endpoint, the interface Yelp documents for conversational, agent-style queries, publishes a response schema whose business object carries name, rating, review count, price, categories, photos, review snippets, and hours. In that published schema, a licenses field does not appear at all.

Nothing here is broken. Every field is doing exactly what its documentation says it does. That is the finding.

The record behind the default

Anatomy of a verification with an undocumented route

The signal exists, and it is already machine-readable. This is not a digitization gap. The public documentation establishes something narrower and stranger: the credential is fully structured, and the interface examined here that carries it, Business Details, ships with the field turned off. What the conversational system gets to see, neither document says.

The gate is a class, not a target. Several partner-tier fields ship default-off. No motive needs to be inferred, and this edition infers none — the mechanism alone is the story. A trust signal that must be requested does not behave like a trust signal. It behaves like an add-on.

What the AI Chat response object visibly exposes is a familiar layer of consumer-facing business information. Ratings, review counts, price, categories, photos, review snippets, hours. It does not expose the verified license as its own field. That documents an output difference, not the full set of signals Yelp may use internally. Filed as Signal Fragmentation, one of the five clinical characteristics of Digital Derangement Syndrome: the authority signal exists, verified and structured, but the interfaces expose it unevenly: gated by default in Business Details, absent as a labeled field in AI Chat’s published response. Whether the conversational system receives or uses it internally, neither page discloses.

Verification is not recognition

“Get verified” has been the standing advice of the platform era, and the businesses that followed it did nothing wrong. The verification is real. What this edition files is the part the advice never covered: verification infrastructure and recognition infrastructure are separate builds, and a platform can construct both while only one of them ships by default. The badge reaches every human who scrolls past it. Whether the underlying record reaches the machine layer — the layer that increasingly answers who should I hire? before a human ever scrolls anything — is a separate question. Among the documents this edition examined, only one interface states its access condition — in one sentence most license-holders will never read. That is why Answer Engine Authority treats consolidation as installation work: a credential counts where machines actually read it, not where it happens to be displayed.

Every claim in this edition can be checked from a chair. Both documents are public: the Support Center page that describes the moderators’ work, and the developer reference that describes what this particular endpoint exposes to an authorized client. Fifty editions into this record, the pattern this specimen joins is familiar: the failure is almost never that the proof doesn’t exist. It is that the proof was filed in a locked room — and nothing on the public record says whether the deciding system holds a key. The open question worth sitting with is not about one platform. It is how many other rooms hold a verification of you — checked, structured, real — behind a door no published page says the answering system can open.

Sources

Yelp Fusion API, Business Details endpoint reference — the licenses field (“List of verified licenses for the business. Access is disabled by default. To have access, please contact us.”) and its five sub-properties — docs.developer.yelp.com.

Yelp AI Chat API endpoint reference — the documented response schema for conversational queries — docs.developer.yelp.com.

Yelp for Business Support Center, “What is the Verified License badge” — moderator verification and paid-feature status — biz.yelp.com. Yelp Data Licensing, AI Agents — business.yelp.com.

Additional partner-tier fields in the same Business Details reference carry default-off access gating — messaging status (is_enabled), Request a Phone Call (rapc), and Request a Quote (raq), their descriptions reading “Access is disabled by default. See Yelp Partner APIs on how to get access.” — docs.developer.yelp.com.

Field-level claims were verified against the reference pages’ published OpenAPI definitions on August 21, 2026 — confirming that the licenses field appears in the Business Details schema but not as a separately labeled property in the published AI Chat response schema.

Your license is verified.

Ask what the machines were fed.

The shield sits on the page, checked and recorded, doing its work for every human who sees it. The two contracts this edition examined tell a shorter story: the credential is structured, gated by default in one, unnamed in the other — and what the conversational system actually holds about you, neither answers. That is not a gap your analytics can close. It is a question for SIA — the Intelligence Officer,
briefed on every edition of this record the morning it releases.

Every edition, in order, from No. 001 · A new edition releases daily, 05:30 CT.

Open the Record
‹ Edition No. 049 Edition No. 051 ›