FOR STATE AND LOCAL GOVERNMENT

Verify the consultant's report before it becomes the basis of public action.

AI now drafts the consultant deliverables, vendor reports, grant applications, and public submissions an agency acts on, and it drafts the orders, rules, and fiscal notes an agency publishes. A wrong date, a fabricated citation, or a miscalculated estimate in these documents does not stay a draft error. It becomes the public record.

The errors are the kind no one catches until they are published: a wrong date in an order, an invented statutory reference in a rule, a fiscal note whose arithmetic does not hold. Asking another AI to check the draft is not a fix. In our pre-registered study of twenty company profiles, three frontier checkers disagreed by as much as 9.6 to 1 on the same document, and the harshest one returned zero errors on documents containing sixteen verifiable ones. A clean self-check from a single model is not evidence that a document is clean. Which checker the reviewer happens to choose moves the error count more than the actual errors do.

The AI is the witness. The math is the judge. GauntletScore verifies each claim against primary sources, then computes the trust score with deterministic math. The agents gather the evidence; no model sits in the verdict. The same evidence always produces the same score.

What GauntletScore checks
Case citations against court records, so an order or brief does not rest on a decision that does not exist or does not hold what it is claimed to hold.
Statutory and regulatory references against the current regulatory text, so a rule or bill does not cite a provision that has changed or been invented.
Fiscal arithmetic with a deterministic math verifier, so cost estimates, budget walks, and revenue projections add up before a floor vote.
Named entities, dates, and events against the public record, so the facts an agency acts on and publishes are the facts.

GauntletScore also runs a dedicated analysis of cause-and-effect claims, testing whether timing, proportionality, and logic actually hold. A fiscal projection or policy justification whose causal story fails the test counts heavily against the document. Sound reasoning counts only weakly for it, because internal consistency is not external proof.

What the first run looks like

Upload the document at gauntletscore.com. Minutes later you are reading the Gauntlet Report: a trust score with a 95% interval, every flagged claim with its verdict and the primary source behind it, and a cryptographically signed, tamper-evident certificate. The error gets caught before it becomes the public record, and the certificate shows the document was independently verified before release. For documents that cannot leave your network, the system can run entirely on local hardware; in our study, the local deployment retained 96% of cloud score quality. In a live production run, the engine examined nineteen claims in one fluent, credible document and returned two debunked against the court record, each with the source that contradicts it. Every voting agent independently recommended against proceeding.

LIMITS

It does not make policy judgments and it does not verify what no authoritative source can confirm. A claim it cannot check is reported as unverifiable, not as false.

Where it fits
Submission review.

Run the deliverable, the application, or the filing against the record before the agency acts on it.

Pre-publication review.

Run orders, rules, fiscal notes, and reports before they are signed, filed, or published.

Records and audit.

Keep the signed certificate as the record that the document was independently verified before release.

There is a deliverable or a draft on your desk right now that is about to become the record. Run it. Either it holds, or it does not, and you caught it before it published.

Analyze a Document
Three free credits. No card. No call.

Validation study pre-registered on the Open Science Framework, March 2026; in progress, manuscript in preparation.