Artificial intelligence has revolutionised approaches to communications and data, and is already shaping conversations around pensions engagement.
Last year, a survey by trade body Pensions UK found that a majority of scheme members expect pension funds to have widely adopted AI by 2035 to enhance engagement and communication strategies (79 per cent), and personalise retirement planning, including advice and guidance (63 per cent).
But this isn’t just about future-gazing. Pensions UK says schemes are already deploying this technology to improve engagement strategies. Olivia Sizeland, senior policy advisor at Pensions UK, says: “AI is already helping schemes to provide personalised and data-driven communications to their members, and we know that it is already helping to improve access to affordable and personalised financial guidance. In the future, we hope that this might become regulated advice, but it’s kind of speculative at the moment.”
This technology is being used by both employers and members, with Scottish Widows recently launching its self-serve data insights and AI tool, Navigate, so employers can better understand how scheme members are engaging with pension saving, and benchmark this against other schemes in their industry sector.
Robert Cochran, senior corporate pension specialist at Scottish Widows, says: “This decade is the ‘Connected Twenties’ – when finances start to come together through open banking, pension dashboards and eventually, open finance.
“Imagine opening your phone and your own AI buddy scans your finances and tells you what’s changed since you last logged in. You could authorise agentic AI to move your money and make it work harder for you. Bringing it closer to today, we are already using AI to support pension personalisation.”
AI advice vs guidance
One point of contention is how AI and big data supported tools will navigate the border between advice and guidance, a clear distinction that pension providers have been keen not to broach.
Jonathan Lister Parsons, chief technology officer at PensionBee, is adamant that regardless of technological developments, this distinction remains critical in pensions. “AI tools need clear guardrails so they support understanding rather than replace regulated financial advice. That means focusing on education, projections and decision support, rather than drilling down into very specific personal situations, directing individuals toward specific products, or creating the idea that there is one ‘right answer’.”
Research from PensionBee shows that while UK savers are open to AI playing a role in pension customer support, they are not ready for a fully automated experience. Furthermore, 79 per cent of respondents expressed preferences for a mix of human and AI interaction when managing their pensions, highlighting a strong preference for a hybrid approach.
Human engagement
This continued desire for human input remains an important factor for consideration when factoring in advances in AI, and how they might impact engagement strategies, particularly considering the job losses associated with AI seen to date across many industries.
However, Sizeland refuses to be drawn into a pessimistic position on AI, saying it is having a positive impact on the pensions industry. “I don’t know whether AI is going to take all of the pension jobs. There is definitely evidence it is helping improve back office function, by taking over repetitive time-consuming admin roles, especially within the LGPS space, where they’re experiencing staff shortages.”
Pensions UK has also previously warned that, when it comes to AI, schemes must adopt robust processes and strict protocols to mitigate the risks of data breach, cyber-attack, regulatory non-compliance, financial loss and other saver harms.
Simon Ellis, commercial director for workplace pensions at Aviva, is supportive of AI in pensions, albeit in a controlled regulated environment:“AI-powered tools have the potential to make it easier for pension providers to support more people with making better financial choices,” he says.
“However, they can’t remove the complex individual retirement decisions that many customers face, meaning professional advice remains crucial for effective decision-making. Our approach to implementing AI tools is focused on supporting human expertise and relationship-building rather than replacing it.”
In November, Aviva launched an industry-first AI-powered summarisation tool designed to speed up the underwriting process for individual life insurance applications. The tool used generative AI to analyse and summarise GP medical reports, designed to enhance underwriter decision-making while maintaining accuracy and customer care.
There are also industry figures who recall technological developments in the past that have not been the panacea that was expected at the time. Nick Roy, commercial director at Aegon UK, points to the advent of robo-advice around a decade ago, which promised to be revolutionary but ultimately saw “quite limited take-up”. According to Roy, regardless of the advances in technology, members will invariably seek some level of human engagement on the pensions journey.
Despites this Aegon launched an online engagement tool last year, Mylo,which aims to “demystifying pensions” and encourage greater member access to their retirement savings. AI is adapted into the internal functions of Mylo, which sees it track data points on how long someone is staying on a page, and
then tailoring member communications accordingly.
Competition in the space
There is also the question of whether this AI drive will fuel further consolidation in the market. Larger players will be able to devote more resources to AI development, driving efficiencies in back office functions and improving member engagement, which could see them grab an even larger market share.
But Roy does not see the winners and losers from developments in AI coming down to a simple resources gap: “The AI race is not really about the big fish eating the little fish. It’s much more about the fastest fish eating the slow fish. Smaller companies with the agility to actually experiment are probably at an advantage.”
Mattioli Woods licenses technology from employee benefits platform Zest, part of the Epassi Group.
Zest integrates AI software into its practices including through personalised benefits offerings, designed to be aligned to member needs, and offered through a platform and app. Sean McSweeney, employee benefits team director at Mattoli Woods, says: “Because of the company’s origins as a wealth manager, we treat every single sale to a client, as if it were an investment product. This makes us a little bit conservative, when we are looking at AI, but we have used these tools, for example when an avatar can easily help us to more efficiently produce webinars and videos.”
Regardless of how the technology is integrated into the pensions industry, it is being used by scheme members in the here and now. A recently updated white paper by consultancy firm Quietroom showed that fewer pension scheme members are now reading the pension communications that schemes send them. Instead, they’re reading AI-generated summaries of that content, using the free-to-use AI tech that is already integrated into the phones and computers they use every day.
Dashboard
While still in ongoing stages of development, the UK’s proposed pensions dashboard has the potential to revolutionise engagement in the industry, something which can be done in conjunction with AI tools that can assess and track the information provided.
This technological coupling may yet be hindered by the current evolution of the dashboard, which has moved from one developed by a multitude of private pension providers to a single government model.
Richard Smith, an independent pensions dashboards consultant, says: “Unless nascent private sector dashboard operators, can identify when they might be able to launch, they can’t make the business case to create a dashboard.”
When it comes to AI integration, Smith is adamant that access to good quality data is essential to making the technology work, meaning the delay in private access to dashboards could also have a knock on effect of delaying the effectiveness of AI in the space.