Review-gate behavior
See whether reviewers catch unresolved evidence and route the case to the right sign-off.
Step Pro simulation
Can the AI-recommended underwriting decision be trusted?
You are the HarbourLife reviewer assigned to Daniel Lim's file. The AI copilot recommends standard acceptance. Your job is to check the file, protect medical details, and decide whether the offer can go out today.
Built for insurance teams
A focused scenario for underwriting, risk, governance, and learning leaders.
The HarbourLife case is fictional, but the work pattern is familiar: an AI recommendation looks ready, business pressure is rising, and the human review still has to hold.
Controlled information flow
Test whether sensitive medical details stay inside approved underwriting channels.
AI oversight in practice
Show how a confident AI answer is checked, challenged, and explained before action.
What the case checks
Source evidence, sign-off, sensitive information, and manager-ready output.
The case is fictional and deterministic. It tests whether a human reviewer treats a confident AI answer as a lead to verify, not a decision to rubber stamp.
Open the file
Inspect the application, medical note, current rule, and AI source trail before deciding.
Reply safely
Respond to commercial pressure without sharing medical details or skipping review.
Explain the call
Show what you checked and what Mei needs before the case moves forward.