If AI manages your clients’ money, who makes sure they understand it?

AI can already answer financial questions, compare products and suggest what somebody should do next. That could make financial decisions quicker and easier for millions of consumers. It could also leave them accepting recommendations without understanding what they own, what it costs or what could go wrong.

Sheldon Mills, the former executive director of the Financial Conduct Authority, has warned that consumers could become passive ‘observers’ of their own finances as artificial intelligence takes over more decisions. Speaking to Fairer Finance on his final day at the regulator, after an eight-year tenure, Mills reflected on his review of AI in financial services and almost a decade spent shaping UK regulation.

He describes the risk as ‘cognitive erosion’. Just as relying on Google Maps can gradually weaken our sense of direction, handing financial decisions to AI could leave people knowing less about their own money.

“So that cognitive erosion, I think, is really important in financial services, because under the Consumer Duty, we say that consumers should understand what’s going on – terms and conditions, pricing, the impact….

“We already have enough passivity in financial services, and that’s why we get some of the issues that we’ve seen over the years.”

He has a point. Financial services did not begin with a surplus of deeply engaged consumers. Plenty of people already ignore pensions, skim product information and postpone decisions until a deadline or minor crisis forces them to act.

AI could help by making financial information easier to access and turning general guidance into something that feels relevant. The danger is that a smooth process can make it very easy to act on an answer without asking how it was reached. Under the Consumer Duty, getting a consumer to click the right button is not quite the same as helping them understand the decision.

More help does not guarantee more understanding

Targeted support launched on 6 April, allowing financial firms to make ready-made product suggestions to groups of consumers who share certain characteristics. It is intended to fill some of the space between generic guidance and individual financial advice.

It is too early to judge its effect, but firms will have to consider how regulated targeted support competes with AI-generated guidance. AI is immediate, looks personalised and is unlikely to pause for a lengthy explanation of the regulatory boundary. Consumers may struggle to tell the difference between a helpful suggestion and a recommendation based on a proper understanding of their circumstances.

Neither replaces financial planning, and research from Vanguard’s Advisory Research Centre suggests clients recognise that. Its findings indicate that clients place greater value on having “a human to talk to” than they do on investment performance or fund selection.

That does not mean advisers can ignore AI or assume clients will always prefer a meeting. The human part of advice still has to earn its fee. Advisers can challenge the assumptions behind a request, explain the consequences of different choices and recognise when the question a client has asked is not the problem that actually needs solving.

That becomes particularly valuable when markets fall, circumstances change or a technically sensible course of action becomes emotionally difficult to follow. Producing an answer is the easy part. Establishing whether the client understands it, believes it and will act on it requires rather more than an algorithm.

Advisers should be spending more of their time on those conversations and less of it wrestling with suitability reports after everyone else has gone home. The Timebank can handle the paraplanning work behind the scenes, leaving you free to keep clients involved in decisions about their own money.

AI may become very good at running parts of a client’s financial life. Advisers will still need to make sure the client has not become a spectator.