A chatbot's answer comes with no redress, and most do not know

Ask a chatbot whether your home policy covers an escape of water from an upstairs bathroom and it will answer you in seconds, in full sentences, with no hedging worth the name.

Then it is wrong, and nothing happens.

On 19 October the Financial Conduct Authority put a number on how little that is understood. Only 40 per cent of consumers correctly recognise that there is no recourse if something goes wrong after acting on advice from a general-purpose AI tool.

The same speech reports that four in five less experienced investors aged 18 to 40 have already used AI to help with investing, and that 86 per cent of firms now use AI themselves.

Set that against the route that does exist. The Financial Ombudsman Service took 4,100 complaints about car and motorcycle insurance between April and June 2026, up from 2,800 in the same quarter a year earlier, and upheld 26 per cent of all the cases it closed.

Here is my problem.

Sixty per cent of people do not know that a wrong answer from a chatbot cannot be escalated anywhere. A regulated firm that gives a customer bad information about a claim owes them a complaint process and, in the end, an ombudsman with the power to direct redress. A free app owes them nothing, and the better its answer reads, the less likely anyone is to check it.

The confidence of the answer is doing the work the accountability used to do.

Our industry benefits from this quietly. Every question a customer takes to a chatbot instead of to us is a question we never have to answer and a complaint that never gets logged. That is not a saving. It is a liability nobody has billed us for yet.