The news: Neptune Flood launched a flood insurance quoting app within ChatGPT earlier this month that enables prospective customers to receive real-time preliminary quotes. It allows them to ask about their flood risk and coverage options before being directed to the insurer’s website to complete the process. Neptune's proprietary underwriting system, Triton, was built as an API-first, modular platform. That architecture made the ChatGPT integration possible without changing its core workflow. A lightweight MCP layer sits on top of Triton to handle data retrieval, risk modeling, and rating in real time within the conversational interface.

Zoom out: Insurance apps for LLM interfaces represent a shift in where insurance discovery happens. As those apps become more sophisticated, they will drive deeper changes to the customer acquisition funnel. Tuio, a Spanish insurer, and Aviva, a UK insurer, recently launched home insurance apps on ChatGPT. Insurify introduced one for auto insurance.

Insurers are not alone in their LLM distribution ambitions. In February, Experian launched an auto insurance comparison app inside ChatGPT that lets users ask questions, review coverage options, and compare estimated rates from more than 37 carriers. Prospects are then directed to Experian’s site to complete the process. Experian’s entry signals that aggregators are staking out LLM distribution too, raising the bar for insurers to build a meaningful presence in the channel.

Implications for insurers: Neptune’s ChatGPT app is an early example of a growing cohort of insurers embedding part of the shopping process inside a widely used AI tool. Conversational AI could become another acquisition channel, especially for products that require consumer education.

By bringing education and quoting into chat apps, insurers can meet prospective customers when intent is high but no purchase decision has been made. The LLM app is especially valuable for complex, infrequently purchased products like insurance, where customer confusion causes friction and human assistance is expensive. But Neptune's architecture made the integration possible without touching its core workflow — an advantage legacy insurers with monolithic underwriting systems won't easily replicate.