Posts

Showing posts from August 30, 2026

Why AI Cannot Fix Unclear Accountability

Image
A large US health insurer managing tens of thousands of workers compensation claims annually asked us to help integrate claims processing, risk controls, and care coordination into a single platform. The integration worked. Information flowed across functions that had previously operated in silos. The platform identified cases requiring intervention earlier than the previous process had allowed, and the recommendations it generated were consistently well grounded in the available evidence. Then a claimant outcome went wrong. Claims explained their role clearly. Risk pointed to the controls they had maintained throughout. Care coordination described the recommendations they had issued at the right time. Every team had done what its function required. The documentation was in order. The process had been followed. Nobody owned the outcome. This is a different problem from unclear decision rights, though the two are related. Decision rights define who has the authority to decide. Accountab...