The wealth management industry is embracing AI, with an increasing number of banks rolling out new ChatGPT-style technologies and tech teams to support them.
[See also: Family offices invest in AI future]
LGT, J.P. Morgan Private Bank, St James’s Place and RBC are among the leading names who are investing heavily in the development of AI systems. Some rely on in-house teams, while others have tapped up the expertise of a growing number of AI-enhanced wealth planning start-ups.
The benefits are numerous – streamlining workflows, reducing costs, improving client experience and better synthesising client data. According to a recent survey by EY of wealth and asset management executives, more than one in five managers believe data ingestion by AI tools will soon be able to support ‘alpha-generating’ strategies – underscoring a sense of optimism by the industry for how far AI tools will soon be able to support operations.
[See also: The FIG regime: A Labour olive branch to non-doms?]
Yet none of this will come without risks. Even giants in the field like ChatGPT experience disruptive ‘hallucinations’ that could be particularly damaging within the private client world. There are also concerns around regulation, privacy and security.
Battle of the banks
Last week, LGT became the latest in a string of firms to announce it had rolled out a proprietary AI tool for employees that provides ChatGPT-like services to ‘summarise documents, extract information and draft texts’.
Content from our partners
J.P. Morgan Private Bank launched its internal generative AI ‘copilot’ tool on its internal ‘Connect’ platform in May, a move the firm has said can save its advisers a couple of hours of administrative work each day. Karen Donnelley, the digital tech leader of the US private bank, has a 50-strong product team dedicated to future developments. BlackRock, meanwhile, announced in December that it was introducing AI ‘copilots’ on its popular Aladdin investment management platform.
St James’s Place, the UK’s largest wealth manager, is developing similar tools. John O’Driscoll, divisional director for business development and advice, tells Spear’s the group has introduced a ‘Advice Assistant’ and generative letter-writing software. There are plans on the horizon for an ‘SJ-GPT’ generative AI tool to support its advisers, retrieving things like regulatory information from across the business.
The deployment of AI tools at St James’s Place is already saving ‘between an hour and two hours per case of just back office time, recording, documenting and setting up the advice file,’ O'Driscoll adds.
These tools are in place for what’s currently a small proportion SJP’s business, including investment advisory, though there are plans to increase this proportion over time.
Outside expertise
The UK wealth management and IFA market has also benefited from enlisting the expertise of AI-enhanced wealth planning startups. These include Planner Pal, a seven-month-old start up which has already attracted 350 customers in the wealth management market. The software allows wealth managers to automatically update records on their Customer Relationship Management (CRM) platforms, fill in rote compliance documents and transcribe meeting notes.
Its founder, Mark Whitcroft, a former director at Deutsche Bank, says AI will empower wealth managers in the near term to complete these ‘low value’ activities quickly, rather than replacing the more abstract and intellectually demanding aspects of their work. ‘People often think about wealth management in a black and white way. But when you look at a wealth manager’s role, it's a job made up of lots of individual tasks. Some are value-adding tasks, and others aren’t,’ Whitcroft says. ‘Research suggests managers can spend up to 70 per cent of time on administrative tasks.’
[See also: First look inside Park Modern’s £60 million trophy penthouse on Hyde Park]
Some other wealth management firms are turning to ‘end-to-end’ tech platforms which, along with AI-powered tools, can simplify operations for their wealth managers. FNZ, whose banking clients include Alliance, Barclays and Quilter, has created an ‘end-to-end’ wealth management platform enhanced by AI to help wealth managers with accounting and reporting, ESG data, know your customer (KYC) checks and regulatory compliance. Its services are currently being used by over 650 financial institutions and more than 10,000 independent wealth manager firms.
‘What our wealth platform brings today is not only modern technology, but also improving and de-risking the economic model,’ says Philippe Bongrand, FNZ managing director for private banking. ‘Piecemeal outsourcing isn’t very effective, but full end-to-end platforms make sense for private banks who want to combine scale and agility in a much more resilient and regulated way.’
Cutting costs, saving time, improving decisions
Such tools will be useful at a time when wealth managers in the UK are feeling the squeeze of increasing costs and pressure from the Financial Conduct Authority to ensure they are providing ‘fair value’ for clients under the Consumer Duty framework. The FCA is also calling on the industry to do more to close the ‘advice gap’ between the relatively small number of UK residents taking financial advice and those that could benefit from it.
'For our advisers, especially those who are part of smaller practices, they’ve got more time to spend with clients, and can service more clients, which can help to close the advice gap,’ O’Driscoll says of the benefits of AI. ‘Most of what I’m focused on is how we use AI to support advisers.’ This tech research includes both ‘natural language processing solutions’ and add-ons which can understand investment industry jargon in human speech.
AI also allows advisers to gain a better understanding of markets by streamlining the process of research and analysis. HSBC, for instance, has refined its AI Markets data and analytics service, allowing wealth advisers to access ‘a wide range of historical data’ along with access to ‘a library of content from HSBC global research’. The firm says its AI-powered natural language processing tool ‘allows users to generate bespoke financial market analytics’ with just a few keystrokes.
[See also: Family offices race to professionalise services amid industry surge]
‘AI is not just a buzzword for us; it’s a fundamental pillar of our strategy for digitising at scale,’ says HSBC's Christiane Lindenschmidt, the chief digital and data officer for markets and securities services. ‘Through thoughtful use of machine learning, we are able to offer our clients more personalised experiences, make better data-driven decisions, and stay ahead of the curve in an ever-evolving industry.’
Katherine Waller, a former RBC wealth manager who co-founded wealth management outfit Six Degrees at the end of 2023, is optimistic that AI tools will be useful for managers to deepen their understanding of clients. ‘For the HNW population, the data set is incredibly opaque and poor. Using AI gives the opportunity to analyse a large data set from a sentiment standpoint,’ she says. ‘It would be very straightforward to pick out trends and themes from client conversations, which could add some real colour to information about the client base and what they’re doing today.’
Waller’s co-founder, Oliver Saiman, highlights the benefits it could provide in profiling these HNW clients, explaining: ‘At the moment, [profiling] hasn’t really changed that much in 20 years — it is effectively a questionnaire. AI could analyse what a client has done in the past, and how they've reacted to various market events. You could then play it back to that client and test their appetite and capacity in a much richer and more useful and personalised way.’
The end of the wealth manager?
Some, however, believe AI will go even further, in some cases replacing the role of managers outright in making investment decisions through the rise of so-called ‘robo-advisers’.
Currently, some retail 'robo-adviser' tools exist which allow investors to learn about different types of investments, including Nutmeg and Wealthify, though investment decisions are made by a team of human experts.
[See also: UHNW clients seek to swap wealth management firms for family offices]
Yet PwC says that in the future, sophisticated tech solutions could soon allow ‘robo-advice to offer the kind of personalised solutions that would once have been reserved for high-net-worth clients, while recognising the need to develop a hybrid human and digital delivery model’. Research predicts the amount of assets managed through robo-advisers will double over the next few years to $5.9 trillion by 2027.
The risks of the AI revolution
Saiman is among those who warn that there are risks and regulatory considerations when such tools begin to offer financial advice.
‘It probably won’t be too long until smaller firms with very high-volume small clients will start to turn to AI, to say “let’s look across our client base, and highlight the number of clients where the financial planning isn’t optimised, and please suggest solutions,’ he explains. ‘Now — if it quacks and has yellow feet, it’s probably a duck — that’s financial advice. So, there will come a time when the FCA will have to take a view on what the lines are between AI providing financial planning advice versus providing insight and information, which is a little bit of a grey area.’
[See also: Family offices raise allocation to equity and fixed incomes despite uncertainty]
In May, along with the Bank of England, the FCA published an update on its stance on how AI is being used in the financial advisory market. It noted that ‘technological innovation can also reduce operating costs, improve efficiency, and more effectively manage risk,’ but added that the regulator would ‘prioritise understanding current deployment strategies within the firms we regulate, so that they – and we – can clearly identify risks and mitigate them.’
Hallucinations are another common shortcoming of existing AI technology – inaccurate information made up by bots which is presented as fact. AI tools also frequently throw up incomplete information, adding a further problem for those looking to the machines for cogent answers. According to research from Stanford University published in May, AI tools used in legal research hallucinated in at least one in six queries. 'Given the high rate of hallucinations, lawyers may find themselves having to verify each and every proposition and citation provided by these tools,' researchers found.
[See also: No gold rush for private client fund managers?]
Yet O’Driscoll says such challenges can be overcome if ChatGPT-style tools could provide more sourcing for the information they generate, which could minimise the risk of fever dreams from the robots.
Another challenge can arise for firms where wealth managers are using openly available AI tools at work – ones that might not have been approved in a regulated environment. ‘We’re quite clear that advisers should not be using generative AI that’s not approved for use in the business,’ adds O’Driscoll.
There are also considerations around client confidentiality. ‘We’re a tech-enabled business, and we’re proud of that. Clients are able to see, visibly, live, their entire financial data, which is really important, because that’s something [private] banks don’t provide at the moment or tend not to provide,’ Saiman says. ‘We realised the biggest risk of doing that was around the security. It’s much easier not to provide that level of tech and do everything in a more analogue way.’
While there will be inevitable issues around client confidentiality when using AI to analyse client conversations, Saiman says he is confident that ‘the wealth-tech market is mature enough now to have got its ducks in order, and to act in a really responsible way around data encryption. Around 10 years ago, that wouldn’t have been the case.’
Still, he says, there are some things clients rely on that an AI may never be able to reproduce. ‘With a HNW market, these can be incredibly complex clients in lots of ways, around their structuring. If you're onboarding a client, for example, having a human who understands that individual and takes a view, and can be flexible sufficiently, is really, really important.’
