My first job out of university was answering phones and meeting clients at an asset management company. I spoke to students starting their first investment, new parents starting education funds for their children, and recent retirees grappling with switching from saving to spending.

I was particularly nervous after being assigned to speak to an economics professor who had taught me mathematical finance only a few months before. I don’t think he recognised me but I sure did recognise him. I took a deep breath, preparing myself for a series of complex questions on portfolio optimisation, alpha and beta, and total expense ratios.

The professor simply asked for an investment, split evenly between the equity, balanced, and stable funds—a curious choice because this combination is just an inferior version of the balanced fund.

Watching my professor struggle to connect his goals to a suitable financial product, and through hundreds of similar conversations, I left the job at the asset management company convinced of one thing: everyone needs financial advice.

Why most people cannot get the advice they need

Firstly, financial advice is exclusive. Quality financial advice requires skilled professionals, and skilled professionals need to be paid. Unless your portfolio is large enough that an adviser’s fee represents a reasonable share of your expected returns, you are not worth their time, and they are not worth yours.

Even those who can afford financial advice run into a second problem: good financial advice is difficult to separate from bad financial advice. Poor advice crowds out good advice because the client cannot tell the difference until the damage is done.

I think of a favourite school teacher of mine, who is approaching retirement now. An adviser persuaded her, years ago, to switch from an investment product into what she believed was another investment product, but turned out to be insurance. She paid premiums for fifteen years on a policy she never needed. When she finally looked to cash out, the money she thought she had been accumulating was gone. This is not an unusual story. You probably have similar stories to tell.

Even if you have access to a good financial adviser, you need to visit them early enough. People tend to seek out financial advice at major life events: a death in the family, a divorce, losing a job. This is the financial equivalent of going to the dentist only when you have a toothache. The routine check-up—unpleasant but preventive—is what most people never get.

Filling the financial advice gap with agentic AI

How should we help the majority of the population that needs good financial advice but cannot get it? Agentic AI—computers that complete tasks when instructed in spoken language—presents an opportunity to provide a new type of personalised financial advice.

The technology has many virtues. It can communicate with people in spoken language—the language of goals and anxieties, not the language of terms and conditions—and translate between the two. It can perform calculations that most people cannot, run through scenarios, and flag mismatches between what someone says they want and what their current financial arrangements will deliver. And unlike a human adviser, computers are available at any hour, can handle many conversations simultaneously, and only costs a few cents for each conversation.

The question is who should provide AI-powered financial advice?

AI companies are already offering financial advice as a by-product of training on the vast archive of personal finance blogs that exists on the internet. For now, this is mostly beneficial to the consumer. But the incentives of AI companies ultimately point toward monetisation, and it is easy to imagine a future in which an AI’s financial recommendations come to resemble advertising more than advice—a platform where the highest-bidding product provider secures the most prominent recommendation.

Financial service providers face similar incentives. A bank that also offers investment products and insurance has an inherent incentive to steer customers toward its own offerings, and to obscure the fee structures that make those offerings less attractive than alternatives.

Consumer financial protection agencies occupy a different position. Their mandate—ensuring that consumers are appropriately served by financial products—is the mandate of a good financial adviser, but without the exclusivity. Also, the agency is not trying to sell financial products. Their incentive is to ensure that consumers are using financial products that fit their needs.

Consumer financial protection agencies also have access to information that no other actor can easily assemble. They can compel financial providers to disclose prices, returns, claim rates, and complaint histories. They can conduct mystery shopping to verify the promises of each financial product. They can monitor complaint patterns in ways that reveal systemic failures before those failures become scandals.

Open banking legislation is extending these possibilities further. In principle, a regulator with the appropriate legal authority could use an API access to check, say, whether a particular individual already holds an active life insurance policy before recommending that they take one out.

Finally—and this may be the most important structural advantage—consumer financial protection agencies can act without waiting to be asked. Financial advisers see clients when clients come to them. But most people never come. A consumer protection agency has both the mandate and the authority to reach out proactively, to conduct what you might call roadworthy inspections of people’s financial lives: checking that the products in place are appropriate, that nothing is obviously missing, that no one is paying for protection they already have or forgoing protection they need.

Think back to the teacher I mentioned earlier. She did not seek out bad advice; bad advice found her. A proactive check—even a simple automated review of what she was paying and what she was getting—might have surfaced the problem years before she did. The technology to do this now exists. The question is who will deliver it.

My argument is this: almost everyone needs financial advice, most people cannot access good financial advice, and agentic AI offers an opportunity to deliver good financial advice—but only if it is provided by an institution with the right incentives, the right information, and the right to act before people know to ask for help.