The pensions industry is in a strong position to broaden its involvement with artificial intelligence (AI), according to experts at PwC.

The firm recently unveiled the latest figures from its Low Reliance Index, which showed that the UK’s 5,000 defined benefit (DB) pension schemes had maintained a £400bn surplus over the previous month.

Given that generous cushion, it pointed out, sponsors and trustees now have “an opportunity to make sure that they reap the benefits” of advancements affecting the pensions industry – including AI and generative AI.

PwC Head of Pensions Technology and AI Gavin Sharma notes that providers are already using AI in their schemes to boost operational efficiency – from automating data processes to drafting scheme documentation. But in the near future, he says, “we can also expect to see AI being used to personalise the member experience and to bring insights that enhance decision-making”.

Essentially, Sharma’s comments herald the dawn of more effective and intuitive services, which sounds like a simple proposition. But in terms of what that could look like in practice, there is significant complexity for the sector to overcome.

Punitive taxation

For ICAEW Head of Personal Financial Planning John Gaskell, the pensions market has taken gradual, but significant, steps away from its initial intent. “As a concept,” he says, “a pension should be relatively straightforward. The key benefit is its tax-advantaged status – meaning that you get tax relief when you put money in. Once that money is secured in the pension, the returns it generates are generally free of tax. And when it’s time to take the money out, you are generally entitled – within limits – to withdraw a proportion tax free. So, the whole idea is that you derive a compound return on a tax-advantaged investment.”

That’s the core concept. But in Gaskell’s assessment, complications have gradually set in as a result of various governments’ changes to UK tax rules. The introduction of the annual and lifetime allowances, for example, put a ceiling on the total amount that people could pay in per year, and the total they could accumulate before their savings were subject to punitive taxation and other complex rules.

More recently, Gaskell notes, a law change that took effect before the recent general election abolished the lifetime allowance. However, he stresses, there are “certain legislative ambiguities around that – along with complications around death benefits.”

Following those changes and others, Gaskell says: “Trying to navigate that landscape is complex and therefore expensive – and quite frankly, as a result, rather off-putting. The main reason that expense has arisen is because complication requires professional understanding.”

Moving goalposts

The impacts of that growing complexity have done the industry no favours, Gaskell says. Providers have frequently had to change their systems and processes to comply with legislative changes, which is costly. And certain products do not necessarily reflect the full breadth of the advantages that can be offered under current legislation because they are legacy items devised under bygone regulatory conditions.

All told, Gaskell notes, politicians, regulators and industry chiefs “have created an environment that’s difficult to understand and in which the goalposts keep moving. As such, an ecosystem of advisers has grown up around the field to help people make sense of it.”

That, though, brings its own share of problems. “The Financial Conduct Authority (FCA) has rules around the provision of advice,” Gaskell points out, “and where the boundaries lie to determine the types of guidance people can give and/or receive without it falling under the FCA’s scope.”

He continues: “Unfortunately, there are grey areas here. And what that does in the minds of providers is set up a notional electrified fence that they’re reluctant to get close to, in case they get a nasty shock. These issues have been festering for decades.”

Going forward, then, Gaskell views the industry’s primary challenge as simplification. And this is where he thinks AI could play a valuable role – not so much by untangling the systemic knots as helping consumers find their way around and through them.

Desired goals

“Where an AI tool could help,” he says, “is with fact-finding and scenario planning. For example, savers plug in data that describes their current circumstances. Then the AI factors in the current legislative environment – a preset that’s constantly updated to capture any changes – and the tool provides a plotter of different pathways or scenarios that will show people various options for how to reach their desired goals.”

Gaskell says that such a tool could also include a ‘probability projector’ that tags scenarios red, amber or green based on feasibility, to help people make fully informed decisions. He says: “Think of it as an automatic gearbox – a device that does something incredibly simple, but has lots of complicated stuff going on underneath that the driver never needs to know about. The algorithm inside the gearbox does all the intricate heavy lifting – and that’s linked to a sort of AI sat-nav and fuel-economy gauge that helps you stay on track and ensure you have the resources to get to where you want to be. And if you change your destination, it just rejigs everything.”

He adds: “The important thing is that it should encourage people to want to get in the car and go on the journey.”

ICAEW Head of Data Analytics and Tech Ian Pay also sees value in AI helping people to steer through this complex terrain – potentially as a chatbot-like tool in a scheme's membership area to answer questions and analyse historic performance.

As pensions are tightly regulated, he says, providers will need to minimise ‘hallucinations’: outputs that look credible and plausible, but are factually inaccurate. In addition, he points out: “Such tools will need to be trained to be aware of their limitations and what they can or cannot say. The last thing we want is for an AI chatbot to give financial advice. So, there must be a clear distinction between a factual analysis of a saver’s historic performance and what that could mean in terms of future investment planning.”