When the car insurance quote is a ChatGPT chat, who owns the customer?

Liberty Mutual started writing auto and home policies inside ChatGPT on February 9, 2026, becoming the first major US carrier to sell directly through OpenAI's chatbot. Months later, the trade has not closed a basic question: when a customer buys through the chatbot, who services the renewal? The candidates are the carrier back office, an independent agent, or no one at all.

The launch landed during an unusual year for the independent channel. First Connect's 2026 State of the Industry Report, cited by Insurance Business America, found year-on-year declines in every major measure of agent-carrier friction, the kind of relationship score that tracks disputes, chargebacks, and back-office handoffs. Three-quarters of carriers still do not use a third-party market intelligence provider to track direct-to-consumer pressure. The relationship between carriers and independent agents is improving on most measures, even as a new distribution surface begins routing some leads around the agent entirely.

Plymouth Rock Assurance followed in late July, opening a home-insurance quoting plugin on the same OpenAI surface. Tuio and Insurify were already among the first hundred apps approved when OpenAI opened the platform to customer-facing insurance flows. The carrier in-chat products lean on the carrier's own rating engine for the actual quote, with the language model handling the conversation. Liberty Mutual has said the model does not set the price. None of the launch materials from any of the four names describe what happens after the policy binds.

A chatbot-originated policy still needs someone to walk a household through an endorsement, handle a billing dispute, take the first-notice-of-loss call, and follow up at renewal. Independent agents, the local brokers who sell for multiple carriers and hold the household as a client, do that work for a living. The current carrier apps in ChatGPT do not specify whether a lead is routed to a servicing agent, assigned after the fact, or closed direct without an agent in the loop. A chatbot-originated book of business has to be owned somewhere, and no carrier has published where.

None of the existing AI rules cover the agent-relationship question. The National Association of Insurance Commissioners' Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted by 23 states and Washington, D.C., governs how insurers use AI in underwriting and pricing. New York's Department of Financial Services requires carriers to explain AI factors in those decisions under Circular Letter No. 7. A national AI evaluation tool is now piloting in 12 states. Those instruments name the rate. They do not name the renewal.

OpenAI counted more than 900 million weekly users as of February 2026, a pool no individual carrier website comes close to matching, and the trade press is treating the new apps as inevitable. The unresolved question is whether the carrier, the platform, the agent, or the regulator owes the customer a defined handoff at the renewal. The first answer will come from whichever carrier publishes a clear servicing flow, or whichever regulator asks for one. None has yet.