To suggest wealth and asset managers have been wrestling with the adoption of artificial intelligence is something of an understatement. As in so many sectors, there are important decisions to be made around where humans can still bring value and where tasks are best outsourced to a machine – and AI brings the potential for real efficiencies. At the same time, there are areas where its impact could be a good deal less positive.
Commissioned by the Financial Conduct Authority and chaired by its executive director, consumers and competition Sheldon Mills, the recent Mills Review into the long-term impact of AI on retail financial services sought to offer some regulatory perspective. It concluded there was significant potential for AI to resolve the advice gap and address consumer inertia on savings, while making it easier to switch between providers.
The report suggests AI could help consumers make better decisions on investments and address the problem that little more than a third (35%) of UK adults hold investments and, of those with £10,000-plus in investible assets, three out of five (61%) have at least three-quarters in cash. “AI could help consumers understand trade-offs, identify when guidance or advice may be needed and help manage savings and investments more actively,” it argues.
So this could be a moment of real potential for wealth and asset management businesses that can harness AI effectively. It could encourage more people to save and invest, while creating a longer-term pipeline of growth. Combined with the regulator’s targeted support regime, there would now appear to be a genuine opportunity to expand the investor base and create a new generation of wealth.
Established use cases
AI also offers advantages in the running of wealth and asset management businesses. One 2026 survey found that four-fifths (81%) of firms were adopting it at some level, with 40% at more advanced stages of scaling or transformation. For its part, the Mills Review found most established use cases are still internal and operational, such as process automation, software engineering, data management and knowledge work.
“For investment managers, AI is enhancing research, streamlining operations, improving client engagement and supporting better outcomes for investors,” says John Allan, director of the innovation and operations unit at the Investment Association. “Combined with innovations such as tokenisation and distributed ledger technology, it can drive significant improvements in efficiency, accessibility and personalisation across the investment value chain – provided it is deployed in ways that sustain trust.”
This is growing increasingly important in certain parts of the market – for example, boutique asset and wealth managers have struggled against a growing onslaught of regulation in recent years, which has raised fixed costs and made it increasingly tough to operate.
According to Jock Glover, CEO of the Independent Investment Management Initiative, AI offers a meaningful opportunity to address this issue. “It can made firms significantly more competitive,” he argues. “A Claude plug-in, for example, can do jobs such as building cashflow models quickly, saving analyst resources. Also, smaller and more nimble companies are generally better at integrating this technology into their business.”
“AI can drive significant improvements in efficiency, accessibility and personalisation across the investment value chain – provided it is deployed in ways that sustain trust.
AI should be built around explainable, trusted and repeatable algorithms with clear, reasoned and repeatable assumptions, where a given set of inputs always produces a given set of outputs.”
Furthermore, says James Sullivan, head of partnerships at Tyndall Investment Management, as AI brings down the cost of basic planning, clients who were previously unprofitable to serve may now be able to obtain the advice and planning they need.
“That is a positive in itself but there is also a democratisation story here – and I think it is the most economically beneficial one right now,” he adds. “The Mills Review highlights that only 9% of UK adults currently take advice each year and only 30% hold life or income protection – that is more than just a ‘gap’.”
Devil in the detail
Nevertheless, there are risks to outsourcing too far – a hint of which sits within the Mills Review’s observation that “consumer journeys” will become “agent-led”. “AI systems will move beyond offering information and recommendations towards trusted AI agents that can act continuously for consumers within agreed limits, providing ongoing financial management and optimising people’s financial lives,” it elaborates.
That sounds superficially attractive – but the devil will of course be in the detail. There is a real danger of commoditisation here, with a handful of funds receiving the lion’s share of inflows. This is already happening to some extent in the MPS space but could become significantly more pronounced if algorithms are involved in fund selection.
“AI comparing products and monitoring portfolios continuously – even executing tasks within set parameters –will inevitably lead to a more commoditised wealth management offering,” says Sullivan. “I would argue, however, that areas such as basic portfolio construction, product comparison and routine reporting were never really where the value creation sat .
“What AI cannot easily replace is judgement, or even vulnerability, in complex or emotional situations. It is a little like decorating a new home – the paint on the walls is what it is, but what makes a house a home are the furnishings, the personal touches. At this juncture, I feel like AI is the wall paint.”
As such, Sullivan uses AI on the process side – for example, analysing funds, writing up meeting notes for investment committees and fund manager due-diligence, setting agendas and ‘sanity-checking’. “What I do not use it for are the judgement calls,” he asserts. “AI speeds up the laborious house-painting – but I still choose the furniture and the artwork.”
Shared reliance on similar models, datasets and infrastructure providers could generate correlated behaviour, herding, opacity and common points of failure across the financial services system.”
The Mills Review does recognise commoditisation could be a problem, noting: “Shared reliance on similar models, datasets and infrastructure providers could generate correlated behaviour, herding, opacity and common points of failure across the financial services system.” AI could, it acknowledges, “enable bias, opaque pricing and personalised manipulation” – leading to the warning: “Consumers will need to be able to oversee, understand and challenge AI-driven decisions.”
The report also highlights concerns around AI agents “enabling continuous personal financial management and execution, leading to automatic and rapid switching”, which could also be a problem for investors. In general, the received wisdom has always been that trading adds costs but little value – that would still be true, even if an algorithm were involved.
Contract of trust
As Dynamic Planner CEO Ben Goss puts it, the more autonomous the AI, the more governance matters. “Good governance will be central to keeping the contract of trust between advice firm and client intact – as well as across the broader industry,” he continues.
“AI should be rigorous, data-grounded and tested under regulated conditions. It should also be built around explainable, trusted and repeatable algorithms with clear, reasoned and repeatable assumptions, where a given set of inputs always produces a given set of outputs.”
In this regard, the Mills Review has made it clear that relying on AI will not circumvent consumer duty. “Outputs can still be plausible without being correct – a tendency commonly called ‘hallucination’ – and firms may not always be able to show how a result was produced or why it should be relied on,” it warns. “In financial services, that can have direct consequences.”
‘AI of the beholder: From process automation to strategic alpha’ – Read more on this here
There are also potential impacts on employment within the sector. Here, Sullivan says he is less worried about staff who sit in front of clients than those one or two steps back – “the analysts, chief investment officers, researchers and support functions whose work is an input into a recommendation rather than delivering it”.
“If you do not control the flow of money, AI can absorb your function far more easily,” he adds. “There remains significant uncertainty about careers within financial services and what the industry will look like in 10 years’ time.”
The Mills Review highlights the potential for AI to solve some of the finance industry’s key challenges and bring efficiencies to wealth and asset managers alike – but businesses need to be clear about the risks inherent in how the technology is adopted. With potential dangers including commoditised solutions, herding and new points of failure in the market, the governance parameters will need to be robust.
