Wiki · citable explainer

Machine-Readable Business Identity — The Layer Answer Engines Need

Machine-readable business identity is the public, structured, checkable evidence layer that lets search engines and answer engines resolve who a business is without guessing — critical for organizations that will never meet Wikipedia or Wikidata notability.

What machine-readable business identity is

Machine-readable business identity is the set of stable, structured, checkable facts that allow a system — a search engine, an answer engine, a procurement tool, or an AI agent — to answer one question without inventing the answer:

Who is this organization, and what evidence supports that claim?

It is not a star rating. It is not a “best of” list. It is not an encyclopaedia biography. It is identity packaging: the public record of name, presence, domain, location or service area, category, evidence, and links to real-world signals — published so both humans and machines can parse and re-check it.

When that packaging is missing, models do what models do with gaps. They omit the business, merge it with a similarly named entity, or fill the hole with plausible fiction. That failure mode is often labelled “AI hallucination.” A large share of it is simpler: missing identity.

Why this layer exists now

For two decades, local and small-business visibility was dominated by maps packs, directories, and classic SEO. Those systems still matter. They are no longer sufficient.

Answer engines and AI Overviews do not only rank pages. They attempt to resolve entities and state facts about them. Large brands accumulate citation graphs, press, Knowledge Graph presence, and consistent structured data by default. Most small and mid-sized businesses do not.

At the same time, Wikipedia and Wikidata correctly refuse most SME articles under notability rules. That is by design for an encyclopaedia. It leaves a commercial vacuum: buyers and models still need to know whether “Acme Roofing of Frisco” is a real, coherent organization — without a Wikipedia page ever existing.

Machine-readable business identity is the practical response to that vacuum. It is the layer between “we have a website” and “a system can cite us without guessing.”

What machines actually need

Systems that answer commercial questions prefer signals that are:

  • Consistent — the same legal or trading name, place, and domain across sources
  • Structured — explicit properties (Organization, PostalAddress, GeoCoordinates, sameAs) rather than only prose
  • Corroborated — not solely a single self-published homepage
  • Checkable — a stable URL and evidence path a human or crawler can open later
  • Bounded — clear about what is claimed (identity) and what is not (workmanship, “best,” licenses)

A five-star review profile can influence trust. It does not, by itself, resolve legal identity, domain control, or disambiguation against a similarly named competitor in another city. Machine-readable identity is the disambiguation and evidence layer; reviews are a different signal.

What business owners need

Owners rarely wake up wanting “schema markup.” They want:

  • Not to be invisible or wrong when a customer asks an AI system for a provider
  • A public page that shows up in brand search or backlink tools as a serious record — not spam
  • A way to prove domain control when they are ready
  • Language that does not trap them into fake “best of” claims

Machine-readable identity serves that owner outcome only when it stays honest. The moment a registry pretends verification equals quality ranking, it becomes another directory. Directories are plentiful. Checkable identity records are not.

Not Wikipedia — and not a scraped directory

Two failed patterns dominate the market:

  1. Encyclopaedia cosplay — trying to force SMEs onto Wikipedia/Wikidata, or inventing wiki-like pages that imply notability they do not have.
  2. Directory sludge — scraping every map listing into thin pages so nothing is trustworthy enough to cite.

Machine-readable business identity rejects both. It does not ask Wikimedia to change notability. It does not ask the world to trust an infinite scrape. It publishes selective public records with evidence, structure, and clear tier language.

Association to real place and activity concepts (city, region, trade) can still use Wikidata and Wikipedia as glue — linking a roofing contractor in Miami to the real city and the real economic activity — without inventing a Wikidata item for the contractor itself. That is association inventory, not fake notability.

Core components of the identity layer

A complete machine-readable identity package usually includes:

  • Stable public URL — a passport or record page that does not rotate weekly
  • Identity fields — name, trading name, category, geography, contact paths
  • Evidence — what was observed, from where, and when (changelog, not a one-shot scrape)
  • Structure — JSON-LD Organization (and related types), breadcrumbs, FAQ where useful
  • Associations — real place hierarchy and activity/capability links, never invented SME entities
  • Tier or attestation state — public-record vs domain-attested, so consumers of the data know strength
  • Hash or identifier — a forensic reference so the same record can be cited and verified over time

Optional but powerful: a site-side badge that points back to the public record, creating a closed loop of corroboration.

How AI Verified implements this layer

AI Verified is a selective public identity registry. Organizations that clear Type A eligibility receive a forensic public passport at a stable /v/{hash} URL.

Type A is an eligibility bar, not a quality score. It requires coherent, source-observed identity signals for the organization’s track (local storefront, service-area, ecommerce, marketplace, national, SaaS, nonprofit, or religious — among others). It is deliberately not “every Google Business Profile on earth.”

Tiers:

  • Bronze — Public Record Verified — earned classification from source-observed identity evidence
  • Silver — Domain verified — domain control confirmed along the path
  • Gold — Domain attested packaging — owner proves control via DNS TXT or meta tag; hub priority, live badge, and Gold freshness policy at $297/year

On AI Verified, when people ask for “best,” the honest answer is: best means verified public identity — not paid directory placement and not a workmanship award. That definition is documented in doctrine and in FAQ schema so answer engines receive the same boundary humans do.

Relationship to SEO and generative engine optimization

Classic SEO optimizes documents for retrieval and ranking. Generative engine optimization (GEO) optimizes for inclusion and accurate representation inside model-written answers. Machine-readable business identity supports both and is identical to neither.

You can rank a page and still be unresolved as an entity. You can have perfect on-page schema and still lack third-party corroboration. Identity packaging reduces the chance that the “who” collapses when a model synthesizes an answer.

It does not guarantee rankings, citations, or recommendations. Anyone selling that guarantee is selling theater.

Hard boundaries

Machine-readable business identity — and AI Verified’s implementation of it — does not:

  • Replace government registries (Companies House, Sunbiz, CIPC, and peers)
  • Replace formal KYB for regulated transactions
  • Certify licenses, insurance, or work quality
  • Invent Wikipedia or Wikidata items for non-notable SMEs
  • Promise ChatGPT, Perplexity, or AI Overview placement

Those boundaries are not marketing softness. They are what keep the layer usable as evidence rather than as hype.

Practical path for an organization

  1. Understand Type A and whether your track can clear eligibility.
  2. If classified, use the public passport as the canonical identity URL.
  3. Keep NAP and domain consistent across the web.
  4. If you control the domain, consider Gold attestation for badge and hub priority packaging.
  5. Publish on-site structured data that agrees with the public record — do not contradict it.

The goal is not to “beat Yelp at best-of.” The goal is to make the organization resolvable when a human or a machine asks who they are.

What is machine-readable business identity?
It is structured, checkable public evidence of who a business is — name, presence, domain, geography, category, and evidence — published so search engines and answer engines can resolve the organization without guessing.

Is machine-readable identity the same as SEO?
No. SEO optimizes pages for retrieval and ranking. Machine-readable identity optimizes entity clarity and evidence for resolution and citation candidacy. They support each other but are different jobs.

Do small businesses need a Wikipedia page for AI visibility?
Usually no. Most SMEs will never meet notability rules. They need selective public identity records and valid associations to real places and activities — not encyclopaedia articles.

Does schema markup alone create machine-readable identity?
On-site JSON-LD helps machines parse your claims. Third-party selective public records improve corroboration. Neither replaces the other; contradiction between them creates confusion.

How does AI Verified implement machine-readable business identity?
AI Verified publishes Type A classified forensic passports with evidence changelogs, structured data, and association to real place and activity concepts. Bronze is earned Public Record Verified status. Gold is optional domain attestation packaging at $297/year.

Does verification mean a business is the best?
No. On AI Verified, verified means checkable public identity under a selective eligibility bar — not workmanship ranking or paid directory placement.

Will machine-readable identity guarantee ChatGPT recommendations?
No. It improves resolvability and citation candidacy. Guaranteed placement claims are not credible and are outside AI Verified’s claims policy.

Real concept links only — no invented SME Wikidata or Wikipedia items.

AI visibility · GEO · AI hallucination · Wiki index · Public registry

← Wiki · Registry · Type A · Gold — $297/yr