AI startups target insurance captives as testbed for full automation

The biggest hurdle isn’t technology but human resistance. Musy notes that captives involve multiple stakeholders—actuaries, auditors, lawyers—whose reluctance to adopt advanced systems forces reliance on outdated tools. Huscarl’s AI, for instance, spends half its engineering effort parsing messy data formats. While AI can flag anomalies (e.g., reserve shifts, unusual claims), judgment remains human. Huscarl’s vision is broader: an AI ‘orchestrator’ tying underwriting, actuarial, and legal decisions. Schmalbach of RYSKEX argues AI should support decisions, not replace them, emphasizing accountability and review. The economics drive interest: AM Best estimates US captives saved $8 billion over five years by retaining premiums. Investors are betting on the trend, with Huscarl raising $5.6 million and Luzern Risk securing $45 million. If AI succeeds in captives, Musy says, it will eventually scale to commercial insurers—though the captive market’s chaos may be its greatest asset.