Insurance for the systems that decide.
Castra Risk is a specialty managing general agent. We underwrite affirmative AI liability for organizations deploying autonomous systems, decision agents, and high-consequence models — in defense, healthcare, financial services, industrial operations, mobility, and the regulated enterprise. We write affirmative coverage where the legacy tower no longer answers.
Cyber excludes AI. Errors and omissions excludes algorithmic outputs. General liability was not drafted with autonomous decisions in mind. As exclusionary endorsements widen across the placement market, deployment accelerates. Our business is the line between those two facts.
We underwrite continuously. A model that was safe at bind can drift by Friday. We price drift, not snapshots — through telemetry, action audit, classification stability, and a dependency graph maintained as part of every policy. Our catastrophe methodology treats correlated autonomy failure as a single peril.
The discipline is old. The risk is new.
Three paths into the book.
General counsel and chief risk officers.
Four affirmative coverage lines — autonomous systems, AI agent E&O, model performance warranty, regulatory defense — written affirmatively, bound by specialty paper, priced against telemetry rather than questionnaires.
CoverageSpecialty and wholesale panels.
Submission requirements, commission summary, panel counsel directory, and quote turnaround. We bind through licensed retail and wholesale brokers. Direct quotes are not offered.
SubmissionTreaty and facultative capacity.
Quarterly disclosure of loss data, concentration analytics, and the catastrophe methodology. Both treaty and facultative participation are open. Treaty memo on request.
Treaty memo