Every AI rollout dies in governance. SENTINEL is the AI coach for the part of the job nobody trains for.
Maya, Jordan, Alex, Donna, Ravi and Priya coach the six functional roles. SENTINEL coaches the seam between them. The cross-role governance agent that watches every artefact your team ships and asks the regulator’s question before the regulator does.
Four lenses. Every artefact. Every coach.
Regulatory risk
SOC 2, HIPAA, GDPR, FDA SaMD, ECOA / Reg B, IEC 62443, EU AI Act, Public Sector Equality Duty — pre-loaded prompts ask the right question before the artefact ships.
Decision risk
Flags kill-switch absences, missing rollback SLAs, undefined success metrics, single-cohort PMF signals, vendor-supplied benchmarks treated as ground truth.
Security & privacy
Prompts learners to think about PHI / PII handling, audit-trail preservation, shadow-IT remediation, model-versioning, third-party data egress — before they hand it to a vendor.
Ethical hot zones
Algorithmic-impact assessment scaffolding, protected-class testing, automated-decision review windows, transparency obligations, dual-use risks.
Three scenarios. Each one ends in an artefact your auditor wouldn’t rewrite.
Reviewing a BA artefact for ECOA defence
A BA submits a fair-lending BRD for an AI credit-scoring model. SENTINEL flags 3 missing artefacts (adverse-action notice template, protected-class slicing plan, audit-trail spec) and writes the regulator-facing one-liner the BA should add.
Reviewing a Release Manager rollout plan
An RM proposes a single-stage canary for an AI fraud-detection model. SENTINEL pushes back: "single-stage canary at 5% misses the SOC-2 control review window — propose 1% / 5% / 25% staged with explicit rollback gates and a kill-switch wired to FP rate."
Reviewing a PO acceptance criteria draft
A PO writes “feature meets stakeholder needs”. SENTINEL rewrites it into measurable, audit-defensible AC referencing kill-switch metric thresholds + rollback SLAs + a regulator-facing success story.
Most AI courses cover “ethics” in one slide. SENTINEL coaches it on every artefact.
Functional courses (Scrum, PMP, ITIL) treat governance as a separate exam. AI vendor courses skip it entirely. LeadAI puts SENTINEL inline: every module, every sim, every team capstone — the regulator’s question gets asked while the artefact is still in draft. For enterprise cohorts, SENTINEL ships with custom rules tuned to your specific regulatory surface (your DPIA template, your SOC 2 control library, your internal AI risk taxonomy).