Independent assurance. Start where you actually are.
Every engagement begins with free triage. We route you to the smallest engagement that answers your question, we put starting prices in public because qualified conversations are faster conversations, and we assess only what we did not build.
| Tier | Service | Starting at | Timeline |
|---|---|---|---|
| 0 | AI Risk Exposure call + triage | Free | 30 min |
| 0 | Self-scored AI Deployment Gate worktoolThe five gates and the scoring key, for your own team to run first. Ask for it on the triage call. | Free | Self-service |
| I | AI Deployment GateOne system, evidence-reviewed. A GREEN / YELLOW / RED decision and a Gate Record built for procurement, a regulator, or your board. | $4,000 | 2–4 business days |
| I | AI Exposure BaselineFinds the AI already operating in your business, grades the evidence, maps the rules that apply in your state. | $3,500 | 1 week |
| I | AI Readiness SnapshotTwelve anchor controls (R1 to R12: policy, inventory, roles, risk, impact, data, vendors, pre-deployment, transparency, oversight, monitoring, audit) against NIST AI RMF and ISO/IEC 42001, with a board-ready report and a verdict of CLEARED, CONDITIONAL, or HELD. Scored under SG-MTH-RS-01 v1.0; criteria published on request. | $5,000 | ~2 weeks |
| II | AI Deployment Gate + Remediation PathThe Gate Record plus the Remediation Path: what to fix, in what order, to clear each Yellow and Red, derived from the Record under SG-MTH-GTD-01 v1.0. You or your builder does the fixing; we re-gate. | $9,000 | 1–2 weeks |
| II | AI Governance Gap Assessment | $18,000 | 4–8 weeks |
| II | Conventional App Gap Assessment (GreenLight) | $15,000 | 4–8 weeks |
| II | Inline AI Vendor Risk AssessmentFor the inline AI control inside your product, the guardrail, filter, classifier, or proxy you bought rather than built: documentation and configuration review across eight dimensions and 32 criteria, with a GO, CONDITIONAL-GO, or NO-GO recommendation. Scored under SG-MTH-VR-01 v1.0; active testing of the control on your deployment is a priced option. | $6,000 | 2–3 weeks |
| II | ISO/IEC 42001 Certification-Readiness ReviewFor organizations that already operate an AI management system: a review of your internal audit (or an internal audit performed under your program, never both for the same organization), an evidence-pack review, and a readiness verdict before the certification body arrives. Scored under SG-MTH-42K-01 v1.0; sold with the Gap Assessment that supplies its Annex A position, never alone. | $20,000 | 4–8 weeks |
| + | HIPAA & Health-AI overlaySnapshot from $5,000; Gap Assessment with Security Rule risk analysis $15,000–$25,000. | $5,000 | 2–8 weeks |
| + | GxP / Life Sciences overlay21 CFR Part 11, FDA CSA and ICH Q10 mapped over a Gate or Snapshot (+$6,000) or a Gap Assessment (+$15,000). | +$6,000 | Adds ~1 week |
| + | EU AI Act, international and U.S. state-law conformance overlaysScored inside the Gate for the jurisdictions your system actually reaches. | Scoped with the Gate | |
| + | Model Risk Management for AIFinancial services: independent validation review of an AI model and its governance, in your model-risk vocabulary. The Gate and the retainer serve as pre-deployment validation and periodic revalidation. | $25,000 | 4–8 weeks |
| ↻ | Lifecycle Retainer ScheduleRecurring re-gates on a risk-tiered schedule, on material change, and after any serious incident, on a schedule derived from your Gate Record under SG-MTH-GTD-01 v1.0. Assessment only; we never operate your controls. | $3,000/mo | Ongoing |
| - | AI Red-Team AssessmentSentinel designs the test regime and assures testing executed by a bench evaluator or your own team; we do not attack your systems ourselves. Scored under SG-MTH-RT-01 v1.0. | Custom-scoped | By engagement |
| - | Senior advisory sessionA fixed-fee half day with a principal: second opinion, board preparation, or a vendor or model review. | $2,500 | Half day |
All work is fixed-fee or clearly banded before we start. No meter running, no scope surprises. We do not issue certificates, we prepare you for the accredited bodies that do.
Starting points shown above are indicative. Final scope and fee are confirmed after the free triage call and scale with the size and complexity of your AI footprint. Pricing effective September 12, 2026 and subject to scope.
We assess. We do not build.
As of September 2026 Sentinel no longer offers ISO/IEC 42001 implementation or a managed AI-governance retainer. A firm that builds your controls cannot independently assess them, and we chose the assessment side. If you need a builder, say so on the triage call and we will point you to the market. We take no referral fee, and we will not assess a control we had a hand in designing.
The AI Deployment Gate
A go / no-go decision before your AI system touches real users, and again on every material change, on a schedule, and after any incident. Every system is scored on the same five questions; consequence, not industry, sets the bar. The decision is non-compensatory: a single Red on any gate is a no-go. You cannot average your way past a missing kill-switch or absent validation.
Fitness for context
Is the intended use, decision scope and limit explicitly declared and bounded?
Transparency & limits
Do you understand the model’s class, mechanism and documented failure modes?
Evidence sufficiency
Is there validation on representative data, with stated confidence and bias testing?
Human oversight
Can a human override, escalate or stop it, with a tested fail-safe?
Monitoring & retirement
Can you keep it honest after launch, control change, and retire it cleanly?
The output is a one-page decision backed by a traceable evidence index. Every gate traces to ISO/IEC 42001, the NIST AI RMF and, where relevant, FDA lifecycle guidance and banking model-risk expectations. Sector overlays (HIPAA, GxP, EU AI Act, U.S. state law) score inside the same five gates, so one Record answers every reviewer. Methodology: SG-STD-002, the Sentinel AI Deployment Gate Standard.
Where most clients start
AI Deployment Gate
One system, five gates, one decision, in days. The fastest way to know whether what you are about to ship can be trusted with real decisions.
AI Readiness Snapshot
Twelve anchor controls against NIST AI RMF and ISO/IEC 42001, each rated on documented evidence under SG-MTH-RS-01 v1.0, with a board-ready report and a verdict of CLEARED, CONDITIONAL, or HELD.
Gap Assessment
The full Sentinel Control Map™: 47 NIST AI RMF subcategories mapped to ISO/IEC 42001 clauses, each control rated Met, Partly met or Not met on documented evidence.
Sector overlays
An overlay is not a second assessment. It maps the controls you already have to the rules your sector adds, scored inside the same engagement.
HIPAA & Health-AI
Health apps, AI features and healthcare websites: the HIPAA core, the AI layer most reviewers miss, and the tracking tech that leaks health activity.
GxP / Life Sciences
21 CFR Part 11, FDA Computer Software Assurance, ICH Q10 and ALCOA+ mapped over your AI controls. Inspection-ready evidence index.
Model Risk Management for AI
Banks, insurers and asset managers: independent validation and effective challenge for the generative and agentic AI the 2026 guidance left out.
EU AI Act & U.S. state law
Conformance overlays for the EU AI Act, the international AI laws and the U.S. state statutes your system actually reaches, scored inside the Gate.
Find out where you stand.
Free, 30 minutes, no deck.
The AI Risk Exposure call: we walk your AI footprint, flag your likely obligations, and tell you plainly whether and where you need help.
Book the call →