There is much debate about the potential impact of artificial intelligence, not just on the financial services industry, but society as a whole.
A common concern is the possibility for it to replace jobs, while the Financial Conduct Authority has said that AI could help close the advice gap.
But some firms in the financial services industry have already embraced AI.
In a survey of CFA Institute members, more than half (56 per cent) said their firms were routinely using AI and ‘big data’ solutions in data analysis. A quarter (26 per cent) said their firms were using them to make decisions.
How is the advice industry already using AI?
In September, digital mortgage broker Tembo announced it had developed an internal, AI-driven mortgage criteria tool.
The tool, named Tembo AI, was developed in-house and uses criteria from more than 30 lenders, content from the broker’s website, as well as data from interactions with clients and lenders to provide answers to mortgage-related questions.
According to Tembo, the tool can answer general questions, such as what a joint borrower sole proprietor mortgage is, and specific criteria-related questions on a lender’s maximum loan-to-value, for example.
Tembo’s chief technology officer Geoff Wright says the firm’s aim was to reduce the amount of time that staff spend on answering complex questions.
“[It has] the ability to drill down further still into the answers, matching the cognitive process an operator goes through as they seek to find an answer to their question,” he continues.
“For example, ‘What’s [the lender’s] maximum LTV? Is that the same for flats and houses? In all regions?’
“Given we’re in a regulated space, we’re hyper-focused on the importance of getting the right answer,” Wright adds. “So the AI is trained to only respond when it has a high confidence it is correct, declining to answer if it does not have this confidence.
“As an added fail-safe, Tembo AI shows you its workings out, allowing human operators to validate how it arrived at its answer.”
The tool is a large language processing neural network, explains Wright, and relies on multiple types of machine learning, including natural language processing and deep learning.
“Natural language processing allows computers to understand, interpret and respond to human language; and deep learning enables computers to learn patterns and representations from large datasets.”
While Tembo AI is currently an internal tool, Wright says the firm will make it publicly available after significant road-testing. “Our focus is on expanding the types of questions that can be asked and the quality of answer returned,” he adds.
The system can learn from each client's compounding data as the years go by, thus improving the advice that it is giving.Bruce Ely-Johnston, AdviceBridgeAnother firm that has incorporated machine learning into its technology is AdviceBridge, an automated digital retirement advice and guidance firm for financial planners.
“AdviceBridge was founded on the premise of using AI to improve financial advice,” says Bruce Ely-Johnston, chief commercial officer.
“Our core algorithm is based on reinforcement learning, which is the class of AI algorithms used by Google DeepMind to achieve above-human level performance in a variety of tasks.
“The AdviceBridge system can learn from each client's compounding data as the years go by, thus improving the advice that it is giving. Much like a human adviser as their knowledge and experience grows.”
How AI is improving productivity
Likewise at Wealth Wizards, AI is not a new concept to the advice technology provider.
“[We] have been experimenting with AI in our products for several years, predominantly focused in the early years on improving adviser productivity and quality assurance,” says Wealth Wizards' chief technology officer Phil McGaw.
“These include use cases such as pre-populating fact-finds in real-time, based on a verbal conversation between the adviser and the client. Also, machine learning models that identify anomalies in advice cases across a firm.”
And this year, McGaw says the provider introduced data insights powered by machine learning.
“These enable our financial services customers to look across advice cases that go through the Wealth Wizards platform, and understand correlations between customer demographics such as generational, net income, home ownership status, etc and get a better understanding of what advice or guidance is providing good outcomes for these segments.
“We appreciate the industry may be sceptical about the value of AI in financial advice. But as long as we continue to make the explainability of any AI solutions at the front and centre of developments, we feel that will provide confidence to our customers to start looking at how it can benefit them.”
True Potential chief executive Daniel Harrison likewise says the principle behind the group’s use of technology is improving interactions between advisers and clients, and that it sees opportunities for AI to support this.
“We’re continuing to investigate new ways of using AI within our live chat function. Using the technology to handle basic queries like password recovery and withdrawals means that clients receive timely responses, reducing waiting times and enhancing their overall experience. It also frees up our advisers’ time.
“We are also using AI to streamline information gathering. So when clients make requests like withdrawing funds but they fail to specify the policy to withdraw from, our AI tech will automatically ask for more information.
“We’re also looking at ways of incorporating AI into compliance processes. While this won't replace our human compliance function, it will act as an initial safeguard.
"For example, we plan to use AI to assess suitability reports against predefined standards, providing immediate feedback to advisers and ensuring smoother submissions that benefit our clients.”
We expect that AI adoption rates will only grow over time.Rhodri Preece, CFA InstituteIn investment management, Rhodri Preece, senior head of research at the CFA Institute, says a general characterisation is that firms are starting to deploy AI across different tasks in the value chain.
“From client onboarding through to research and analysis, portfolio construction, trading and execution, there are applications across each of those components of the investment value chain.”
But he also says that AI is not yet widespread in the industry.
“I would say most firms are not yet at that point where they’ve embedded these processes into their business models. I think it’s really more about experimentation and adoption and targeted-use cases, rather than it being a widespread part of how firms do business at this point.
“But we expect that will only grow; the adoption rates will only grow over time.”
Chloe Cheung is a senior features writer at FTAdviser
