AI Audit Exchange
Go back

Guide · By the Ingenious AI team · Published 14 November 2026 · Reviewed 14 November 2026

Auditor Credentials: Claims vs Verified Facts, Explained

What is claims vs facts?

The claims-versus-facts distinction on AI Audit Exchange separates auditor-stated information — credentials recorded as the auditor states them — from platform-recorded facts: registration status under the vetting workflow, engagement matches, grants and room records. The platform's terms state the same division for evidence: auditor credentials shown in the directory are recorded as stated; vetting records the platform's own onboarding decision, not a state licence.

Why the platform records claims instead of rewriting them

When a platform 'verifies' a credential by restating it in its own voice, it quietly takes ownership of a claim it cannot actually underwrite — and readers relax exactly as much as that false ownership invites. Recording claims as claims is the honest architecture: the auditor's words stay the auditor's, attributed and unedited, and the platform's own voice is reserved for facts it can own — statuses, matches, records. The reader is never confused about who is speaking. This mirrors the evidence model in the rooms, where imports carry their producing product's badge and direct uploads are badged as the auditee's own filings. Attribution, everywhere, over assertion.

Layer one: reading the claims

Read credentials as a professional narrative to be explored, not a score to be ranked. Note specificity: regimes named (SB 1119, CPPA), evidence types described, engagement shapes sketched — specificity is testable in conversation, and auditors who write specifically usually welcome being tested. Note also scope honesty: a practice claiming two regimes deeply reads differently from one claiming everything briefly. And note what is absent: a claim set silent on method says something in a marketplace whose whole structure is method — rooms, manifests, seals, scoped grants. Your brief's fit questions then do their work: claims give you the threads; pull the ones your engagement depends on. The questions are laid out in choosing an independent auditor.

Layer two: reading the platform facts

Against the claims stand the facts the platform itself records. Vetted status: the registration passed the owner-run workflow and is listed. Match records: this auditor was accepted for this engagement, on this room, and the grant that followed. Room records: the work done under engagements — artifacts examined in rooms that retain their question threads and access logs. Facts compound differently from claims: an auditor's second engagement with you rests on the first engagement's room record, which neither side can embellish after the fact. When you weigh a new auditor, you have claims plus vetting. When you weigh a returning one, you have a record. Price the difference.

Testing claims fairly: the fit conversation

Fair testing is precise, job-related and the same for every candidate. Ask which regimes they have examined under and what the evidence looked like. Ask how they sample — and listen for manifest, sampling and hash verification in the answer. Ask what they need beyond the room; a strong auditor's list is short because the room model is designed to be sufficient. Give claims a chance to become demonstrated understanding, and let vague answers stay vague on the record of your decision. What fairness forbids is the opposite failure in both directions: accepting claims as facts because the profile looked polished, or discounting specific, testable claims that survive testing because scepticism felt safer.

The same discipline, turned on auditees

Auditees live under the mirror-image rule, and recognising it makes you better at reading auditors. Your room's claims — 'our evidence is complete' — are assertions; your room's facts are the manifest, the seals, the validation runs, the retention setting, the access log. The auditor reads your two layers exactly as you read theirs. Organisations that understand this symmetry stop polishing claims and start maintaining facts: the filing rhythm, the worked stale flags, the exported manifests. In a marketplace where both sides' facts are recorded and both sides' claims are attributed, trust stops being a personality trait and becomes an emergent property of the records. That is the Exchange's whole thesis, stated in one paragraph.

Red flags the distinction exposes

Three patterns deserve instant scepticism. An auditor whose marketing elsewhere upgrades 'vetted' into 'certified' or 'state-approved' has failed the first honesty test the platform sets — vetting records a listing decision, not a licence, and the platform is not affiliated with any state registry. A profile whose claims cannot be connected to any testable specifics — no regimes, no method, no engagement shapes — is asking to be read as fact while supplying only adjectives. And any party, on either side, who wants the record-keeping loosened 'just for this engagement' — evidence by email, access without a grant — is asking you to exit the system that makes their claims checkable. Decline, politely, and stay in the rooms.

Five actions to take

  1. Read every profile in two layers: the auditor's claims, and the platform's facts (vetted status, match records).
  2. Test claims with the same fit questions for every candidate: regime experience, sampling method, needs beyond the room.
  3. Let returning auditors be weighed on their room records, which neither side can embellish.
  4. Apply the mirror rule to yourself: maintain your room's facts so your claims need no polish.
  5. Treat any upgrade of 'vetted' into 'certified' or 'state-approved', by anyone, as a red flag.

Worked example

A fictional auditee, Wrenfield Consumer Brands, compares two respondents to its SB 1119 brief. Auditor A's claims are broad — many domains, many superlatives, no regime named. Auditor B's are narrow: SB 1119 examinations for consumer-facing deployers, method described as manifest-first sampling with hash verification, one engagement shape described concretely. Both are vetted — that platform fact is equal, and it is all it says.

In fit conversation, B's claims convert: asked about sampling, B describes weighting decision records across quarters and verifying seals against the manifest — the room's own vocabulary, used fluently. A's answers stay general. Wrenfield accepts B, and the reason recorded in its own notes is the two-layer reading in action: equal facts at the gate, unequal claims under fair testing. Six weeks later the engagement closes with 17 sampled artifacts and 17 verified hashes — a room record that will speak as fact in B's next fit conversation, with anyone.

Frequently asked questions

Does the platform check auditors' credentials?

Credentials are recorded as the auditor states them. The platform's own act is vetting: the owner-run decision to list the registration. That decision is recorded as what it is — an onboarding decision, not a state licence or a credential guarantee.

What on a profile can I treat as fact?

Vetted status, and — once engagements happen — the records the platform keeps: matches, grants, and the rooms' own contents. Everything else is the auditor's stated claim, to be weighed and tested.

An auditor calls themselves 'certified' because they are vetted. Is that right?

No. Vetted means listed after the platform's workflow. It is not a certification, licence or registry status, and the platform is not affiliated with any state registry. Treat the upgrade as a honesty signal — a negative one.

How do I test a claim about regime experience?

Ask what the evidence looked like in that regime's rooms: sources, sampling, freshness issues, close-out. Genuine experience answers in specifics that match the room model; borrowed claims answer in generalities.

Do auditees have claims and facts too?

Exactly the same structure: your assertions are claims; your manifest, seals, validation runs, retention setting and access log are facts. Auditors read you in two layers just as you read them.

Keep reading

All AI Audit Exchange guides · Auditor directory · About AI Audit Exchange · Security · Pricing · AI Audit Exchange home · Previous: AI Auditor Vetting: How the Owner-Run Workflow Works · Next: How Matching Works Between Auditees and Auditors

Multi-Client Evidence Rooms: An Auditor's Practice Guide · Evidence Rooms for AI Audit Agencies and Teams · AI Audit Exchange Plans: Rooms, Retention and What Each Tier Includes