AI in finance: how is technology changing financial advice?

Financial advice is an incredibly important part of managing your money, but it is inaccessible for many people. Artificial intelligence (AI) could be ready to change that.

Financial advisers view AI as an opportunity rather than a threat: Benchmark Capital, Schroders’ financial advisory arm, found last January that three quarters of the UK’s financial advisers expect to start incorporating AI into their advice process within the next five years.

How soon would you anticipate incorporating AI based technology applications in your advice process in some way?

Subscribe to MoneyWeek

Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE

Get 6 issues free

Sign up to Money Morning

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Sign up

So-called ‘robo-advisers’ or digital wealth managers have become a popular form of simple financial planning that is widely available for most users.

“AI is already transforming the financial services landscape, and as it continues to evolve it will fundamentally reshape how consumers interact with their money and financial products,” says Phil Turnpenny, policy executive at The Investing and Saving Alliance (TISA). “By encouraging innovation, the transformative power of AI can empower millions to make better financial decisions and have a bright financial future.”

According to the Bank of England, 75% of financial service firms are already using AI and 10% more plan to adopt it over the next three years.

Use cases are varied, but one area of focus is the potential for AI to bridge the advice gap, democratising access to financial advice.

AI in financial services

AI clearly has a big role to play in financial services. If we include more traditional forms of AI like machine learning (ML) or natural language processing (NLP) in our definition, then AI has been used in finance for decades.

Generative AI (genAI) is a newer phenomenon, at least as far as everyday usage is concerned. In terms of public awareness it dates back to the launch of ChatGPT in November 2022. This is the kind of AI you’re likely most familiar with; it can generate text, images or videos in response to specific inputs (“prompts”).

There’s an important distinction between the two. While traditional forms of modelling in financial services leverage varying types of “deterministic” models (meaning they typically return the same precise answer every time), genAI models are “probabilistic” in nature. GenAI based responses are the model’s best interpretation of the most probable answer to a specific prompt.

“Hallucinations” are an inherent property of genAI models. It isn’t designed to repeatedly answer questions in an identical fashion, but to generate novel responses to a particular prompt based on the model’s training data.

There is, then, an obvious level of risk when it comes to applying AI to finance. GenAI in particular isn’t well-suited – at least, not as well-suited as many people think – to providing accurate information.

"A useful rule of thumb when using generative AI for ‘high risk’ use cases in financial services: 'if you need to fully explain and repeat the output, don’t use genAI," says Emily Prince, head of analytics and chair of the AI working group at London Stock Exchange Group.

It is, however, very good at certain things. Typically, these are lower-risk tasks, like summarising documents or consolidating multiple information sources.

Done properly, though, AI can help to bridge the “advice gap”, by making financial advice more widely available.

Why would I want an AI financial adviser?

Financial advice, sadly, isn’t available to everyone. There are big barriers to access, and while these are primarily related to wealth, that also means that secondary factors like age start to show big discrepancies.

Schroders’ financial adviser survey 2024 found that 74% of financial advisers will only advise clients with a minimum of £50,000 in assets. More than half (51%) will only advise clients with at least £100,000, and nearly a quarter (24%) will only advise clients with more than £200,000 in assets.

Financial advisers’ minimum asset size for new clients

The average person has £16,232 in savings, according to the Building Societies Association. That doesn’t include other forms of wealth that might be included in that £50,000 level, and it’s a little tricky to compare accurately, but according to ONS data the average Brit aged 35 to 44 has an average total net worth of £195,612. More than half of this consists of property or physical wealth, which most financial advisers wouldn’t include in the calculation.

In other words, half of financial advisers won’t consider working with the average person below the age of 44. That creates an obvious age disparity. Schroders’ survey found that just 7% of advisers have an average client age of 50 or below.

Average age of financial advisers’ client base

All in all, a lot of people are cut off from financial advice. M&G Wealth Advice’s 2023 Wealth Gap report found that just 11% of adults have paid for financial advice in the last two years.

This isn’t financial advisers being stingy. The cost of servicing a client is, on average, £800 per year. If an adviser charges 1.5% of assets, that implies £53,333 in assets just to break even. Given the inputs required, it just isn’t economical for advisers to serve lower-wealth clients – yet these are the people who need financial advice most of all.

How is AI used in financial advice?

The hope is that AI can help to bridge this advice gap by bringing down the cost of providing financial advice.

Financial advisers are already incorporating plenty of genAI to improve their efficiency. According to an upcoming survey from Benchmark Capital, Schroders’ financial advice arm, 45% of firms are already using it.

The most common use case by far is for recording and summarising meetings, but common use cases also include drafting suitability reports, responding to clients, and even identifying clients who might be emotionally or financially vulnerable, but who hadn’t disclosed it.

“I’ve also heard of AI being used to train younger advisers coming through,” Gillian Hepburn, commercial director at Benchmark Capital, tells MoneyWeek. “The AI can record a meeting between a more experienced planner and their client, and one between someone more early in their financial planning journey, and then map the two together to identify where the junior planner is not asking the same type of questions.”

AI is also being built into external-facing financial products. Many of these are already making decisions for us: according to the Bank of England, 55% of AI use cases in financial services involve a degree of automated decision-making.

One good example is robo-advisers. These offer ready-made portfolios based on risk appetite, as well as a degree of AI-led portfolio management (like automated rebalancing). Some popular providers are InvestEngine, eToro’s Smart Portfolios, IG Smart Portfolios and Moneyfarm.

Some companies go further, though. Octopus Money, for example, is setting out to democratise financial advice using AI, by making a holistic service more widely available.

“Most financial services companies have chosen to just focus their efforts on elements of personal finance, instead of the entire process of optimising all aspects of somebody’s finance,” says Tom Francis, head of digital advice at Octopus Money.

Octopus Money’s platform – currently in testing, due to launch in May – is a hybrid model: a highly-trained human adviser is still doing much of the work, particularly everything involving directly interacting with the customer. But much of the heavy lifting and analysis, as well as recommendations, is executed by technology.

That cuts down the amount of time advisers have to spend with any given client, freeing them up to do the more human-specific work. This cuts costs, and allows financial advice to reach more people.

“There are a number of factors that make democratising financial advice really difficult to execute. Technology is well suited to overcome some of them – but not all of them,” says Francis.

How accurate are AI financial advisers?

If you’re taking financial advice from AI, it’s important to understand exactly how accurate the source that you’re using is.

In general, steer clear of unspecialised generative AI platforms like ChatGPT.

A study from Investing in the Web found that ChatGPT answered one in three questions about finance and investing incorrectly – not something you’d want to happen if you were trusting it with your investments.

Not only do genAI tools often give wrong answers, but they frequently don’t have access to the latest information on things like share price movements.

“ChatGPT search is a welcome application that can save us a lot of time, but it can also misdirect us if we’re not careful while using it,” says Pedro Braz, CEO of Investing in the Web. “Where the subject matter is as serious as personal finances, overuse of AI can have a significant negative impact.”

If you are going to use ChatGPT for financial advice, do so relatively early in the process, says Braz, “to help clarify terms or shed light on a general topic.

“Using it to ask more personalised questions is where the risk of miscommunication comes into play,” he says, adding that it is also a good idea to ask your generative AI tool for a list of sources so that you can fact-check its response.

Specialised financial advisers ought to have accuracy checks built in. Octopus Money, for example, uses a complex algorithm – in essence, an enormous decision tree – to generate its recommendations.

“This ensures that the advice we give is consistent, free of bias and scalable, so that anybody (regardless of wealth) can access it,” says Francis. “We place a huge emphasis on helping customers create accurate and detailed financial plans, which our digital logic can then interpret.”

Francis adds that tech checks are conducted against customer data monthly and after every tech update, ensuring that the algorithm is functioning as intended.

The upshot of it all is this: AI can provide sound financial advice, but you should only trust this advice if it comes from a bespoke, industry-specific source. Out-of-the-box genAI platforms like ChatGPT may not be accurate enough to be relied upon.

The need to regulate AI-generated financial advice

The ability of genAI to make mistakes means that there are obvious caveats to its use in financial advice. Industry bodies warn that there needs to be regulatory oversight of the technology in order to protect consumers.

“Not all AI is created equal,” says Turnpenny. “Search engine-based AI tools are increasingly being used as a source of financial advice – often without oversight, transparency, or accountability.

“This has created a wild west situation and poses serious risks, particularly for vulnerable consumers who may be misled by confident-sounding but inaccurate information.”

TISA is campaigning for the Financial Conduct Authority and His Majesty’s Treasury to bring genAI search engines within the purview of financial regulation in order to protect consumers.

“Our research shows that properly designed and regulated AI has the power to improve financial access and outcomes – especially for underserved groups,” says Turnpenny. “We must draw a clear line between risky, unregulated tools and those that are built with consumer protection at their core.”