The recent approval by OpenAI of the first insurance-focused ChatGPT app, designed by US intermediary Insurify, has sent shockwaves through insurance markets.
Following the announcement, major US broker stocks such as Willis Towers Watson, Aon, and Arthur J. Gallagher declined by roughly 8% to 11% on February 9, reflecting investor anxiety over AI’s disruptive potential.
Investors interpreted the development as a potential threat to traditional brokers and intermediaries, triggering sharp sell-offs in related stocks. However, industry experts have called the market reaction “overdone,” noting that the technology is more likely to augment and partner with traditional distribution channels rather than replace them outright.
Berenberg insurance analysts Michael Huttner and Carl Lofthagen argue that while AI applications like ChatGPT’s new insurance capabilities raise concerns about distribution networks, they also offer substantial operational benefits to incumbents that could counterbalance this risk.
Unveiled this month, Insurify’s app enables users to access tailored rate estimates for auto insurance through conversational prompts and then complete purchases via its own platform. The tool integrates detailed personal and vehicle data to generate personalized insurance comparisons.
Rather than disintermediate brokers, analysts predict that AI platforms are more likely to partner with intermediaries who provide regulated reach and trusted customer relationships, which AI platforms alone cannot fully replicate.
Berenberg analysts pointed out that the regulatory burden and liability exposure of selling insurance directly are significant hurdles that OpenAI and others may not want to manage independently.
“AI is likely to work with aggregators and brokers to sell policies, helping them grow rather than replacing them,” Huttner and Lofthagen told markets.
Trevor Jones, a partner in the insurance practice at advisory firm West Monroe, echoed this view.
He highlighted that AI is already being used behind the scenes to streamline producers’ work, including making data navigation, risk assessment and quote generation more efficient, which in turn could enhance broker productivity.
But while AI can increasingly handle standardized or simple coverages, trust and expertise remain decisive in complex risk areas where personal advisors add demonstrable value.
Jones said brokers are likely to demand integrated flows and value-added data services in exchange for aligning with AI distribution platforms.
“Whether that’s data benchmarking, access to customers, or otherwise,” he said, “there is a compelling enticement for artificial intelligence organizations to bring some of the larger brokers on, both to learn from them and to have them as captive on their channels.”
As AI platforms gain traction, concerns about data privacy and governance are also rising. Insurance data encompasses highly sensitive personal and business information, and Jones warned that questions about the use and access of such data on AI platforms remain unresolved.
“It’s well known that OpenAI ingests a lot of information from around the world. It’s also well known that there are a lot of ethical and moral concerns about that,” noted Jones. “On both sides of the equation, there are problems that need to be solved.
“I don’t think insurance companies are going to give OpenAI or Anthropic discretion to use their data, and absolutely not their customers’ data. However, when you’re sharing via a platform and submitting risk – whether it’s a personal line or perhaps even a business line – absent agreements, audits, compliance, and regulations, that data is certainly going in the direction you think it will.”
