Crosby to Insure AI Agents for Professional Liability

Crosby, a NewMod law firm, will purchase professional liability insurance for its artificial intelligence agents so they can deliver legal work without human review. The announcement by CEO Ryan Daniels positions the firm to treat its AI systems as autonomous legal associates, a sharp departure from the industry norm of keeping a licensed lawyer in the loop for every output.

The move arrives as the legal technology market experiments with insurance products that cover AI-generated advice. UK proptech company Orbital secured underwriting for AI outputs in residential conveyancing, and startup Corgi launched a general AI insurance offering. Crosby's initiative goes further: it insures the agent itself, not just the output. By doing so, the firm implicitly asserts that its algorithmic workers carry a status comparable to that of a human lawyer, a concept that, if accepted by regulators and insurers, could redraw the boundaries of legal practice.

Daniels outlined that Crosby's lawyers currently review every work product, but rapid improvements in agent performance have made continuous human oversight unnecessary. The firm plans to work with auditors to define evaluation benchmarks and is engaging bar associations, state regulators, and insurance carriers to shape the framework. No insurer has been named publicly. The goal, Daniels said, is to allow clients to receive agent-generated results "as is" or connect to Crosby's platform and accept answers directly, turning the AI from a tool into a service provider.

Industry Implications

Crosby's bet signals that early movers in the NewMod space are ready to structure liability around algorithms rather than individuals. If the model gains traction, traditional firms may face pressure to match or risk being undercut on speed and cost. Insurance companies will have to underwrite risks they cannot fully predict, testing whether actuarial models can accommodate fast-evolving, opaque systems. For clients, treating an agent as an independent practitioner raises knotty questions about duty of care, conflict checks, and malpractice attribution, all currently anchored in human conduct rules.

DreamLegal Perspective

Legal technology vendors should watch how insurers price and condition these policies, as the underwriting data will eventually inform product design and risk tiering. Legal operations teams evaluating agentic tools need to ask whether their current engagement letters and indemnification clauses are compatible with AI-delivered advice that carries its own insurance backstop. The Crosby experiment may accelerate a broader regulatory conversation; any company building autonomous legal agents should be preparing its own compliance narratives now.