[SPONSORED CONTENT]
The Pensions Regulator’s (TPR) recently published AI plan provides a practical blueprint for action: govern with discipline, monitor with care, and keep members’ interests at the heart of every decision. The Mercer Master Trust (MMT) is already adopting this approach to ensure AI supports retirement choices, enhances understanding, and complements human advisers.
Adoption underway with member needs in mind
As the TPR plan notes, AI is becoming a regular starting point for pension advice and support with decision-making. Delivering high-quality, trustworthy information is therefore essential. Many people turn to mainstream AI chatbots for pension answers because these tools feel accessible and reassuring. But that trust is often misplaced, because these tools do not have access to scheme-specific information.
The TPR plan recognises this reality and treats it as a reason to accelerate progress, not to pause. To translate urgency into action, the MMT has brought AI development in-house, as it pilots a member-facing AI pensions assistant tailored for MMT members. The intention is to keep AI use transparent, well-governed, and aligned with savers’ best interests, while preserving rigorous human oversight and accountability.
Understanding, not jargon
What members want from AI is straightforward: clarity and a no-judgement approach. Around one third of those who would rather receive financial advice from AI than from a human adviser cite fear of being judged as the reason.1 AI doesn’t rush or criticise. It simply responds. When AI is trustworthy and well-governed, it can act as a patient guide.
Managed well, this could support better outcomes.2 Digital tools can reduce the cost of providing support and advice, enabling scalable access to high-quality information 24/7. In a world where instantaneous answers are expected, AI has real potential to assist, but the pensions industry must act quickly, because members are already using these tools. Unlike authorised financial advisers or providers, general-purpose AI tools are unregulated. Members’ use of these raises risks around the inaccuracy of information, liability for decision-making and how the advice/guidance boundary applies.3
A focus on outcomes, not the latest development
The TPR plan has a strong emphasis on outcomes, clear understanding, accessible insights, fair access, and robust risk controls. Technology will continue to evolve, but the core outcomes should remain the focus. By concentrating on outcomes, the industry can test, learn, and adapt without being hostage to every new tool. This approach resonates with Mercer, who are focused on developing and piloting an AI pension assistant that delivers accurate answers relevant to the Mercer Master Trust scheme and grounded in expert content. By carefully curating and limiting the content the tool can reference, it significantly reduces the likelihood of hallucinations and incorrect answers. The aim is to deliver reliable, understandable support that helps members plan with confidence.
Governance and monitoring, the essential foundations
Governance sits at the core of a responsible AI programme. It’s not a tick-box exercise; it underpins design, deployment and ongoing monitoring. Rigorous testing and independent assurance ensure AI behaves as intended and remains under appropriate human control. Monitoring is a cornerstone of safety: real-time oversight helps identify deviations, biased inputs, or outputs that could mislead members.
Mercer mirrors this approach with substantial investment in risk monitoring and control with clear triggers for human intervention to create a safe and transparent member experience. Ongoing testing is embedded in operations with pension experts continually evaluating the AI pension assistant and logging results for improvement. A dedicated internal system monitors against misuse and attempts to compromise it, while a library of test inputs and expected responses tracks and maintains quality over time.
Data privacy, security and scam risk
Member trust depends on data privacy and security. As AI capabilities expand, so do risks of misuse, impersonation and data leakage. The plan’s data considerations align with the Mercer Master Trust’s stance: protect member data, limit exposure, and design systems that reduce scam risk. Data should be used to improve understanding and decision-making, not to increase a member’s risk footprint. Members do not get this level of privacy by using mainstream AI chatbots. All AI is not the same.
Turning the plan into protected member outcomes
Ultimately, the way forward is safe, trusted AI that informs and supports members. The Mercer Master Trust is already well advanced on the path set out in TPR’s AI plan, and is embedding rigorous governance, continuous monitoring, and transparent data practices to help members plan for retirement with confidence.
To read more articles from Mercer visit the content hub on Corporate Adviser – here.
____________________________________________________________________
1 Brits worried about ‘being judged’ prefer to use AI for money advice | Personal Finance | Finance | Express.co.uk
2 Review into the long-term impact of AI on retail financial services (The Mills Review) | FCA
3 https://www.thepensionsregulator.gov.uk/en/document-library/corporate-information/ai-plan