Harbor emerges from stealth with $10M seed round to insure AI agents

AI agents are getting jobs at banks and insurance companies. The question nobody has fully answered is who pays when one of them makes a bad call. Harbor, a San Francisco-based insurtech startup, thinks it has an answer. The company launched publicly on October 13, 2026, with a $10 million seed round to build a platform that manages AI agent risk and backs it with dedicated insurance. What Harbor is building Base10 Partners, Abstract, and 8VC led the round. Restive and Innovation Endeavors also participated. The pitch has two halves. The first is a risk management platform, and the second is insurance coverage built specifically for AI agent failures. It starts by building an inventory of the AI agents a company is running, which is harder than it sounds once agents start multiplying across departments. From there, Harbor assesses each agent’s operational, security, and regulatory risk in real time. If an agent is about to make a decision that breaks compliance rules, the platform is designed to step in with automated interventions before the damage lands. The monitoring doesn’t stop after setup. Harbor says its platform reassesses risk whenever the underlying models, tools, or dat...