Groundwork
Enter workspace
AI trust diagnostics

Evidence beforeauthority.

Bring something to evaluate. Groundwork routes it to the right diagnostic and returns a verdict, the patterns behind it, and what to fix first.

Maps findings to EU AI Act and NIST AI RMF. Flags issues for legal review. Never certifies.

What you get back

Sample
Platform Risk
Workflow automation vendor, support tickets to billing
High risk
1 failing · 3 at risk · 2 passing · 6 dimensions
  • Incident Historystrong evidenceFail
  • Data Handling Posturelimited evidenceRisk
  • Compliance PostureinferredRisk
Fix first
Incident History, fail

Obtain the vendor's incident response and remediation plan for the disclosed breach before routing customer data through it.

Six diagnostics

3 live · 3 coming next
Coming next

How it works

03
  1. 01
    Submit

    Give it a vendor, a use case, or a launch plan. It goes to the right diagnostic.

  2. 02
    Score

    Every finding carries an evidence tag: strong evidence, limited evidence, inferred, or no signal found.

  3. 03
    Fix

    One report. A verdict, the patterns behind it, and what to fix first.

Evidence, not vibes

A vendor claim and independent verification are not the same strength of evidence. Groundwork never treats them as if they are.

Vendor claim

“Customer data is encrypted at rest and in transit.”

Vendor security pagelimited evidencePass
Independent verification

“Encryption at rest and in transit confirmed in the vendor's current SOC 2 Type II report.”

Third-party audit reportstrong evidencePass
Read the methodology

Who it's for

For AI product teams, consultants, and procurement leads who need evidence before they sign off.

Built by Nelly Pourmehr, AI UX Lead.

Have something to evaluate?

Start an auditPrivate beta

Groundwork is in private beta. Request access