Pension transfers are a minefield for advisers but the makers of a new tool believe their algorithms can provide the answer
The recent launch of Wealth Wizards’ artificial intelligence capability, Turo, illustrates the possibilities technology can bring to advice. Turo enables robo-advice to be delivered in line with an advice firm’s way of doing things. As a result, it can be applied to more complex areas of advice such as DB pension transfers.
Advice firms input a representative sample of around 20 to 30 DB transfer cases and Turo breaks those decisions down into 18 economic and psychological factors.
“The sample of representative cases gives the computer an idea of what decisions you make and how you make them. This is passed to AI algorithms which figure out under what circumstances DB pensions should be transferred and which factors mean they shouldn’t,” says Wealth Wizards’ chief executive Andrew Firth.
“It“It can also be used as a compliance tool, identifying cases where the adviser may want another look because the advice given may not be expected.”
Commentators describe DB transfers as a “hot potato”, so could AI provide the key to consistent and compliant advice in the area?
Portafina managing director Jamie Smith-Thompson accepts the potential upsides of AI, such as increased efficiency, greater consistency of decisions, a higher output from a given number of staff, easier reporting and checking. However, he sees these as benefits only if costs are reduced and passed on to the client.
“A key limiting factor [of AI] is how you engage with the client to truly understand what it is they are trying to achieve and why – and to understand how they react when you discuss the possible consequences of their decisions,” he says.
“How does current AI pick up on the subtle nuances that tell you
that your client isn’t totally comfortable or entirely sure of themselves? How does it empathise with a client’s situation?”
Progeny Wealth director Alex Shaw says the more that can be done to ensure advice firms have a robust DB transfer process the better.
“I don’t see AI as replacing what we do; it is aiding what we do. We can have the AI data arrangements but still manage human interaction, how the client feels about it and what the family member wants to do. A lot of that can only be gleaned by asking questions over a few meetings. I’m not sure a computer is going to do that.”
FinanFinance & Technology Research Centre director Ian McKenna highlights the progress IBM’s Watson system has made in the treatment of cancer as an example of the power of AI.
“Watson is already sophisticated enough to build a machine comparable with the most sophisticated financial adviser if it wanted to,” he says.
“That is not to say there isn’t a hugely important role for humans in advice, but we need to evolve and work with the technology. The application of AI to identify patterns and behaviour within advice firms has got huge potential. Imagine if the FCA had a version of Wealth Wizards’ system so that advice could either pass or fail?”
A compliance tool
McKenna says, going forward, there could be a system into which firms could drop every piece of advice to identify any potential inconsistencies, biases and risks, thereby reducing compliance risk.
“In the long term it should improve standards and outcomes, identify behavioural biases and ensure firms don’t apply biases in the recommendations they are giving.”
TCC group managing director, advisory services, Andy Sutherland points out there is often confusion between robotics – effectively automation without the application of judgement – and AI, underpinned by machine learning that has the capacity to develop a form of cognitive judgement through continual learning.g.
F the benefits of machine learning from a compliance perspective. Identifying DB transfer cases for review can be done in a fraction of the time using AI relative to humans following a paper trail.
“For example, where the customer objective for transferring their DB pension is accumulation – so no immediate need for taking an income – there is inherent risk that requires additional consideration. This includes looking at factors such as other sources of retirement income the client has got, the shape of their future income needs and their level of investment knowledge and experience.
“AI can tell you if all these factors are present, where they are and whether they are at the level expected. This has potential to save firms significant time and reduce costs,” says Sutherland.
“A further benefit is that when you are running AI over a population of files it can check them all, rather than just a small sample, at relatively immaterial additional cost.”
Sutherland says that once you can teach a machine to identify the most at-risk files, you can then devote your human resources more effectively and efficiently. However, he acknowledges the process needs to be right at the start. Indeed, the process is only as good as the information you put in.