FOR FINANCIAL RESEARCH AND COMPLIANCE

Verify the research note that reached your desk before it moves money.

AI now drafts the research notes, earnings summaries, pitch books, M&A memos, and executive profiles that land in front of a decision. The document was written by someone else's model, it reads with full confidence, and SEC enforcement has already targeted overstated AI claims. A fabricated figure in research your firm relies on or distributes is a supervision problem with your firm's name on it.

The errors are confident and specific: an executive who does not hold the role, a settlement that never happened, a margin figure the filing contradicts, valuation arithmetic that does not compute. Our validation study ran on exactly this document type, AI-generated company profiles. 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. Among the documented catches: fabricated executives, an invented nine-figure corporate settlement, research and development spending overstated roughly four-fold, and plain arithmetic errors in financial walks.

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
Financial figures and executive identities against SEC EDGAR filings, so the numbers and the names in the note match the record.
Regulatory references against the current text, so a compliance memo does not cite a rule that has changed.
Valuation and projection arithmetic with a deterministic math verifier, so the walks, ratios, and models compute.
Corporate-event timelines against the record, so the M&A history and the causal story built on it hold.

GauntletScore runs a dedicated analysis of cause-and-effect claims, testing revenue-to-margin attribution, event ordering, and proportionality. A thesis whose causal chain 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 fabricated figure gets caught before capital or a client moves on it, and the supervision record shows the research was independently verified against primary sources. 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 produce investment advice and it does not verify what no public source can confirm. A claim it cannot check is reported as unverifiable, not as false.

Where it fits
Pre-reliance review.

Run the note or the memo against the filings before capital or a client moves on it.

Supervisory procedures.

Make verification the standing, documented step for research the firm receives or distributes.

Deal diligence.

Check the target profile and the deal memo against the record before the meeting.

There is a note or a memo in front of you right now that money is about to move on. Run it. Either the figures hold, or one does not, and you caught it first.

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.