The Protocol of Deterministic Search

The End of Probabilistic Discovery

For twenty years, local business discovery ran on guesswork. Search engines indexed fragments from directories, review sites, social profiles, and websites, then assembled a best guess for each business. Inconsistent signals. Unverified data. Often contradictory.

The taxonomy behind that era was equally shallow: categories, keywords, ad groups, and service pages. Built for indexing. Never for understanding.

That model worked when the interface was ten blue links. Users absorbed the ambiguity. They clicked, compared, and decided for themselves.

AI does not work this way.

When a customer asks an AI assistant for a roofer in Tampa, the system resolves that query against structured, verified data. If that data does not exist, the AI hallucinates an answer or declines to recommend.

Neither outcome serves the business.

The Deterministic Requirement

Deterministic search means the system produces the same correct answer every time. Three properties are required:

First: structured data. The business defined in a machine-readable schema. Canonical fields, verified values. Not scraped fragments.

Second: verified provenance. Data traceable to the business owner. Not inferred from third-party aggregation.

Third: temporal currency. Data that reflects the current state of the business. Not a cached snapshot from an unknown date.

The IdentityRecord Resolution

IdentityRecord establishes a canonical identity record for each business. Not a profile. Not a listing. A structured data object that satisfies all three deterministic requirements.

The record is built through a resolution process: observed from public data, claimed and verified by the owner, then enriched with structured signals AI can query directly.

The result: AI resolves business queries against verified, owner-controlled data. Deterministically.

The businesses that get structured first get recommended first. That is the protocol.

Local AI Taxonomy

The old local-search world organized businesses around categories, keywords, ad groups, and service pages. That taxonomy was designed for indexing. It helped businesses show up in a list. It was never designed to explain why one business should be chosen over another.

AI needs something deeper. It needs to understand why a business deserves a recommendation.

Local AI Taxonomy captures the drivers behind actual buying decisions. What the business does. Who it is right for. What makes it credible. What makes it different. What reassurance its customers need. What proof supports its claims. What context shifts a recommendation from tentative to confident.

Categories tell a machine what kind of business it is looking at. Local AI Taxonomy tells a machine whether to recommend it and why.

Every module in the product is built on this taxonomy. 33 Truths structure the business around it. The FAQ Engine surfaces questions that map to it. The Proof Engine generates signals that support it. The taxonomy is not a feature. It is the foundation.

Trust Signal Generation

AI cannot access the most important data about a business.

Licenses, insurance, certifications, and compliance records stay private, locked in filing cabinets and scattered systems. AI models are left to infer credibility instead of verifying it.

IdentityRecord solves this by converting private business data into structured, machine-readable signals. The documents stay private. The signals go public.

The Identity Locker

The Identity Locker holds the sensitive data: licenses, insurance certificates, compliance records, professional credentials.

Documents are structured, timestamped, and tracked. IdentityRecord extracts verification signals (valid, expiring, expired) without exposing the documents themselves.

Expiration tracking prevents decay. When a license lapses, the signal updates automatically.