Last week, the FCA announced plans for a live AI testing service designed to make it easier for financial services firms to harness the power of technological innovation. The new service will allow firms to collaborate with the FCA while they check their new AI tools are ready to be used.
Research and consultancy firm NextWealth and has been conducting research to gauge the adoption of AI across the financial services space so far, following the release of its AI Index report last year.
NextWealth’s managing director and founder Heather Hopkins (pictured) explained that, as of December 2024, 35% of advice professionals out of 340 surveyed were using AI in their business. Since then, she expects that figure to have increased, but it’s still fairly low.
Of course, the heavily regulated nature of financial services has contributed to some hesitancy, but actually NextWealth’s research shows that a more pressing barrier to adoption is the difficulty of integrating AI tools into existing tech stacks.
See also: FCA AI service will ‘enhance client trust’ in the technology
“This is an age-old problem,” said Hopkins. “Advisers have got their platforms, they’ve got their back office systems, cash modelling, their portal, their own Microsoft suite of products, and this is just another system that’s going to sit alongside. Firms aren’t necessarily looking at what new technology they can take on because their existing tech stack might be a bit of a mess.”
She noted there are firms working on solutions to this problem, including extracting data from existing systems, but small advice firms with just a few employees don’t have that kind of resource in their business. They’re sticking to using AI in basic ways, such as to record and summarise meetings.
Large firms who do have the money to invest more heavily in AI face their own challenges though, because they must ensure consistency in how advice is delivered in multiple offices. They can’t always be nimble in implementing new solutions.
The problem of data
Another barrier to widespread AI adoption is data availability and connectivity. This could be client-level data such as valuations, and costs and charges, as well as firm-level data such as the percentage of assets going into in-house products, or assets versus market growth.
“That doesn’t sound particularly sophisticated, and it’s not, but there’s a large consolidator I’ve been talking to that is absolutely delighted they’ve been able to get 75% of the client data they need. That’s crazy – how is that good in this day and age? They should have access to this basic information about client assets and costs and charges.
“We’re actually doing a piece of work now to try and understand which providers are doing a really good job and which are providing the data that advice firms need to be able to run their business more efficiently, because we think that’s the sort of thing that could really unlock adoption of AI within advice,” Hopkins said.
The world’s largest asset manager, BlackRock, recently asked shareholders in its American Income trust to approve plans to use AI in portfolio decision-making. As part of the new systematic strategy to improve performance, the trust would use AI, big data and BlackRock’s human expertise, the company said. Shareholders gave the proposals the green light.
Will we see more asset managers introduce AI into decision-making in a more formal way and, if so, does that mean some fund managers could be out of a job?
See also: FE fundinfo acquires AI platform Lunar AI
Hopkins said there is an argument that AI could cause a resurgence in active fund management because it can pick up on subtle cues. These might include tone of voice or facial expressions in response to questions asked on an earnings call. This could be useful input to help a human fund manager run a portfolio.
Hopkins likened it to self-driving cars which will be able to anticipate that a pedestrian will cross at a zebra crossing up to two metres before they get there.
“A driver wouldn’t pick up those signals, but the person may move what they’re carrying to the other hand, they’ll start to look over their shoulder, they’ll take the hand of the child next to them, so the self-driving car has more inputs that it will process,” she said.
“It’s not the replacement of the human but the ability to collect data that maybe before was not available, that could inform decisions. Whether that results in better portfolio performance remains to be seen, but it’s always good to have more inputs than less.”
The AI pay gap
What are the consequences of being left behind in the AI race? Hopkins warned there is already an AI skills gap which is linked to pay disparity. PwC’s 2024 AI Jobs Barometer found that lawyers with AI skills command a 27% wage premium, while Oxford University calculated a 23% wage premium for AI skills, surpassing the 13% premium for people with a Masters degree.
There’s a huge shortage of advice in the UK, so if firms don’t want to embrace AI, it’s probably fine for them. But it could mean missing out on efficiencies and, to some degree, fun, argued Hopkins.
“It makes the job of financial planning more fun. How can you get rid of some of the more mundane, repetitive tasks that maybe advisers don’t enjoy that much? It’s not just about ‘how do we deliver cheaper service?’ or ‘how do we reach clients with smaller portfolio values?’. It’s also ‘who wants to sit on hold to a provider for 30 minutes?’. If you could send your AI agent out to collect that data for you, that would be better.
“For firms that are ambitious and looking to grow, we’ve had case studies at the NextWealth AI Lab about using AI to remarket to dormant leads. It’s not about replacing the adviser in the equation, in the room with the client, because we all know that is incredibly powerful and important. That is why human, face-to-face financial advice has persisted despite a number of societal, regulatory changes over the last decade.”
Making advice more fun
Importantly, it can also improve the client experience. Hopkins said clients usually come out of planning meeting excited and empowered. What follows can be a tedious process of filling in forms, repeating information, documentation being lost, things needing to be faxed, identification photocopied, until six weeks later the client has lost all enthusiasm for the process.”
It’s remembering that there’s an opportunity not to disrupt that human interaction that is so meaningful, but to try and smooth out the bumps in the road,” Hopkins said.
“AI could help keep clients engaged, make the job more interesting, bring down costs, and hopefully get beyond 8% of the population having access to advice. The benefits to society would be enormous.”
Hopkins’ practical tips for getting started with AI
- Just try it and see where you get to. Have the discipline of always trying to use AI for different tasks.
- Whatever your job, think about how you could use this technology to help you work better.
- Share use cases with colleagues in the industry, and discover ways to evidence positive ROI.
- Ask AI for clarification or elaboration, don’t always accept the first answer. Treat AI like a colleague who is very well trained but has no context on the industry or the job. You wouldn’t take something from a junior colleague and then, if it wasn’t right, just throw it away. You would go back and try to iterate, work with it to get to where you need to be.
This story was written by our sister title, Portfolio Adviser