Business Problem & Solution

A marketplace runs on policy: how refunds are handled, what listing content is allowed, how guest data is protected, what partner KYC and rate-parity require, how security incidents are disclosed. Those rules need an owner, a version history, and — crucially — a way to check whether the ecosystem is actually following them. Compliance & Policy Management is the central place to author and version the policies and to sweep ecosystem activity for violations.

It proves a baseline-plus-overlay policy model (owner edits bump a version) and a deterministic compliance sweep that reads real ecosystem activity — refund decisions and listings — and surfaces pass / warn / violation findings with a reason. It simplifies by running rule-based checks rather than an LLM (the surface is AI-assist, not metered), over synthetic data; every finding is labelled synthetic.