AI assistants now answer the questions your buyers used to search — and they paraphrase original research without credit. Facts & Data turns the numbers only you have into registered, machine-readable claims: a stable ID, a timestamped first appearance, dataset lineage, and fact pages built for the crawlers behind ChatGPT, Claude, Perplexity, and Google.
Your buyers ask an assistant and read one synthesized answer. What makes it into that answer is whatever machines can parse, ground, and attribute — and most original data is published as prose that gives them nothing to cite.
Your benchmark travels through blog posts, decks, and AI answers. The number survives; your name doesn't. Nothing marks you as the citeable origin.
A stat buried in a PDF has no stable ID, no dataset, no license, no timestamp. Retrieval systems can't attribute what they can't parse — so they don't.
Google is retiring ClaimReview rich results in Search — the consumer fact-check category dies with it. The first-party side of the claim vocabulary — Claim, firstAppearance, Dataset — is wide open. Own it.
Four stages, one spine: the claim registry. Everything else either feeds it or flows from it.
Highlight a number on a page you own and the extension drafts the claim — text, source URL, timestamp, author. A required-fields gate tells you exactly what's missing before it can be called verified.
A JSON-LD registry per domain: canonical wording, first appearance, the dataset each claim derives from, owner and approver, an append-only changelog, and a computed originality grade.
Compiled fact pages with embedded schema.org markup, sitemaps and recrawl pings, and verified crawler access for GPTBot, ClaudeBot, PerplexityBot, and friends — plus an MCP endpoint that serves agents ground truth directly.
Every claim is fingerprinted and tracked across the web and AI answers — each appearance classified attributed, unattributed, or mutated — and piped into VIZEO for visibility reporting.
Registration isn't a badge — it's a record any third party, human or machine, can audit: a stable URL and anchor, a timestamped first appearance on your domain, and a lineage that terminates in a real dataset with method and license.
firstAppearance timestamps the claim on your domain — the public stake that this number originated with you.
No dataset, no methodology, no license — no verified status. We record what you attest; we never adjudicate truth. The chain is checkable without trusting us, or you.
Every edit appends to the changelog — old value, new value, revision date. Retracted claims say so; they never vanish.
{
"@type": "Claim",
"@id": "…/state-of-ops#claim-automation-roi",
"text": "73% of teams shipped faster within
one quarter of adopting automation.",
"author": { "@id": "yourcompany.com/#org" },
"firstAppearance": "…/state-of-ops-2026",
"isBasedOn": {
"@type": "Dataset",
"name": "State of Ops survey · Q3 2026",
"measurementTechnique": "survey, n=1,204",
"license": "CC BY 4.0",
"distribution": "state-of-ops-2026.csv"
},
"okf:verificationStatus": "verified",
"okf:originalityTier": "first-party-measured"
}
Compiled into your fact page's markup — the page anchor and the record's @id are always identical.
We crawl your domain and extract every factual assertion you've published — research, pricing pages, case studies.
Each claim is traced to its source and graded for originality — first-party measured, derived, licensed, third-party, or unsupported.
You see which claims you can't back up — and which facts nobody can cite you for. The registry starts from there.
The side panel does the whole loop: scan a domain for claim-like statements, capture and grade each one as a record, and audit how complete the registry's provenance chains are.
Crawls up to 25 pages (sitemap first, robots.txt honored), inventories claim-like statements, and maps them against your registry.
Captured from FaaStr — Your AI Co-Founder — filed as publication #1.
How much of each record's provenance chain is filled in — never a measure of whether claims are true.
100Suggestions are generated on this device only — the claim text never leaves your machine. Every suggestion is a chip you apply and attest.
Every engagement begins the same way: we find out what your domain can actually claim. Keep the registry running month to month, or let us run the whole program.
The entry point — every engagement starts here
The ongoing layer — extension, registry & monitoring
The program — for teams that want it run for them
Facts & Data makes your claims citeable. VIZEO, its companion extension, proves they're being cited.
Creates the citeable asset: registered claims, datasets, fact pages, markup, the MCP endpoint, and an originality grade for your whole domain.
Measures whether the world consumes and credits it: appearance reports, citation share, and SEO/GEO visibility — including where AI answers cite you, or don't. Meet VIZEO →
The Authority Report pairs your originality grade from Facts & Data with citation and visibility metrics from VIZEO — one picture of whether your domain owns its data, and whether the world credits it.
Start with a claim inventory audit. We crawl your domain, extract every factual assertion, grade each for originality, and hand you the gap report — the claims you can't back up, and the citations you're leaving on the table.