Twelve questions, ~60 seconds. The result tells you what an AICPA peer-reviewed auditor running a SOC 2 + AI engagement would flag — and the AI-specific controls your existing SOC 2 program is probably missing.
The specific controls that pulled your score down — these are exactly what a SOC 2 auditor or an enterprise customer's vendor security review will ask for when AI is in scope.
Same score, but with: every gap mapped to its specific Trust Service Criteria sub-point, suggested control wording for your auditor's matrix, and a 14-day remediation order. Sent to your inbox in <5 minutes — single transactional email, no spam.
Each question is weighted by the audit consequence of getting it wrong. Failing a question on per-inference logging (CC7.1) costs more than failing one on model card disclosure — because the former blocks an unqualified Type II opinion, while the latter is remediable in the management-response window.
The thresholds reflect emerging AICPA AI assurance guidance plus what we have observed large enterprise procurement teams flag in vendor SOC 2 reviews where AI is in scope. They are calibrated to what the auditor checks, not what a marketing page claims.