The Review That Has to Hold Up: AI Decisions Under Audit and Appeal

​When an AI assisted decision gets challenged, someone has to show what happened. What the system saw, what it recommended, what the reviewer saw, and why they decided the way they did. Most teams find out too late that they cant.

​Eugene Chan (Behavieural) and Elena Rodin (Kombocode) walk through what breaks on both sides. The system side: what has to be captured at the moment of decision so it can be reconstructed later. The human side: how to show a reviewer acted reasonably given what was in front of them.

​We will work mostly from healthcare examples, coverage, prior authorization and eligibility, since that is where the audit and appeal pressure is loudest right now. But the same test lands on any regulated decision an AI touches. Lending, insurance, hiring, legal work, anywhere a person has to answer for a machine assisted call.

​Bring a real scenario, we will take a couple live if time allows.