ChatGPT will drive a new wave of insurance commoditisation
Posted:
The Race to the Bottom Has Already Begun for Consumer Insurance
On February 9th, 2026, insurance went live inside ChatGPT. Within hours, billions were wiped from broker valuations. The market panicked. Analysts called it an overreaction.
They're wrong, just not in the way they think.
The real threat isn't that LLMs will replace brokers, especially brokers of more complex risks like WTW (who lost 15% of their share price in the panic). It's simpler and more brutal than that. AI can only compare on price, and that's exactly what it will do.
We're already seeing it happen. MoneySupermarket's (Mony) share price tells that story, it also fell 15% with the arrival of Insurify and Tuio apps in ChatGPT. It recovered slightly when MoneySupermarket announced their own app. But the recovery was temporary and it's now dipped back down as the announcement of Aviva's direct chat app signalled where insurers are heading. LLMs are destined to become aggregators on steroids, brutally efficient at finding the cheapest option - whether that's through insurers providing prices through MCP services or through an ever-improving ability to agentically source insurance prices. Either way, insurers will go direct and a race to the bottom on price is the inevitable direction of travel.
Why is the Commoditisation of Consumer Insurance Inevitable?
In the UK consumers are conditioned to comparison based on price. The vast majority of consumer insurance is sold via price comparison websites. There's no expectation of a need to actually understand what policies offer and how cover applies. This creates a cycle of dissatisfied customers and regulatory intervention that wouldn't exist except that no-one really understands the insurance they're buying, assuming that it's all vanilla and largely the same. Though seemingly beneficial to customers the fixation on competing on policy price had unintended consequences: insurers building sophisticated pricing engines to acquire new customers at a loss and depending on making profit from putting up prices for renewing customers (forcing the regulator to step in with new regulation to stop non-switchers from being penalised), the stifling of innovation (when price is the sole measure of value, the only reliable way to compete is to cut costs - innovation tends to be expensive), and ever-thinner policies and lower volumes of successful claims. One of the easiest ways to cut costs is to insure less for the same premium. To give a sense of this, for single trip travel insurance the FCA reported that in 2024 only 34% of premiums were paid out as claims. The remainder will go to operating costs, acquisition costs (price comparison website fees) and profit.
Going back to the problem. Cheap is appealing to a customer, but cheap is the wrong answer for insurance.
The cheapest policy is rarely the right policy. A consumer who buys on price alone, with an AI able to hunt across the whole market for the very cheapest policies, will end up underinsured, badly covered, or holding a policy full of exclusions they never understood. When the claim comes, the damage is done.
More than ever, people need someone to understand their actual needs and match them to coverage that genuinely protects them. LLMs can't do this - even if they pretend that they can. To list the things that an LLM can't do:
- recommend specific policies tailored to the customer
- provide insurance advice
- complete insurance applications
- guarantee compliance with the law or regulations
LLMs will focus on what they can do - find prices and show what's cheapest - and bluff the rest.
The role that customers need the LLMs to fulfil, which they can't, are the province of a broker. But the traditional broker model, time-consuming, inconsistent, expensive to scale, can't survive in a world where ChatGPT quotes in seconds.
This is the problem Prompted is solving.
Prompted turns LLMs into instant, robust brokers. Not price-comparison engines, actual brokers. The platform uses data and AI to identify what a user genuinely needs, maps those needs against hundreds of semantically structured policies, grades each policy against the user's specific situation, and condenses that analysis into clear, actionable guidance.
Not the cheapest. The right policy at the best price.
This is where insurance distribution is going. LLMs will lead consumers in a race to the bottom, but that doesn't need to be the case. Prompted is building the infrastructure to make LLMs more than a price engine and turn it into a genuine adviser. For consumer insurance and small businesses, who are underserved by cheap, generic policies of varying and unclear value.
The future of insurance isn't cheaper. It's smarter. The right policy at the best price.