The Pensions Regulator’s Corporate Strategy 2026 to 2031 sets the supervisory tone for the next five years, and for the first time puts the governance of AI in workplace pensions on the regulator’s named agenda alongside value and security.
In reading this article you will understand:
- What tPR’s Corporate Strategy 2026 to 2031 sets out as its core value and security priorities for workplace pension schemes, and how these connect to the wider consolidation and value for money agenda.
- What changed, and what stayed the same, between the strategy’s May 2026 consultation draft and its July 2026 final version.
- What tPR now expects of trustees, scheme managers, administrators and their advisers on the governance of AI, and where the detail is still to come.
The Pensions Regulator’s new five-year strategy, published on 14 July 2026, does something its predecessors didn’t: it names artificial intelligence as an area it intends to supervise, alongside the pensions system’s two long-standing preoccupations, value and security. A Sustainable Retirement Income for All: Our Corporate Strategy 2026 to 2031 was published alongside a detailed Corporate Plan and a Regulatory Roadmap, and together they set the supervisory tone through to 2031. For corporate advisers, that matters before any individual rule change lands. It signals what The Pensions Regulator (tPR) will focus on, and press schemes and employers about, over the period a typical scheme review plays out. Around 23 million people now save into a workplace pension through automatic enrolment, yet tPR estimates 15 million working-age people remain off track for an adequate retirement income. Closing that gap safely, including safely through new technology, is the strategy’s organising idea.
Value and Security: The Framework Behind the Headlines
tPR frames its mission around three pillars, protecting members’ money, enhancing the pensions system and supporting innovation, and translates them into six outcomes, three for members and three for the market. For members: savings that are secure and delivered as promised; better value through investments and services; and fair access and opportunity. For the market: well-run schemes with effective governance and skilled trustees; a resilient market supporting UK growth; and a seamless journey from joining a scheme through to taking a retirement income.
tPR is shaping that market, not just policing it. Its chair, Emma Douglas, who took up the role on 1 July 2026, has said tPR is “moving towards a system of fewer, larger, well-run schemes, able to invest in diverse assets in the interests of members, and potentially the UK economy.” The strategy commits tPR to using the value for money framework, which it is still shaping with the DWP and the FCA, to accelerate consolidation where standards aren’t being met. That direction is covered elsewhere in this series. The point to hold onto here is that AI governance runs on the same track: one more marker of whether a scheme is well-run enough to stay independent, or better placed inside a larger, more heavily governed structure.
What Changed Between May’s Draft and July’s Final Strategy
tPR consulted on a draft strategy from 11 May to 8 June 2026 and received 34 responses across six themes, including calls for stronger cross-regulatory coordination, sharper focus on decumulation risk and clearer treatment of AI. The core six-outcome framework held. What tPR floated in May is substantially what it confirmed in July, so advisers who read the draft needn’t unlearn anything.
Three things did sharpen in response to feedback. The vision statement’s sustainability language now makes explicit that a “sustainable” retirement income is one that lasts through later life, within a system that is itself stable and resilient, addressing stakeholders who wanted adequacy named alongside sustainability. The Prudential Regulation Authority and Pension Protection Fund were added as named regulatory partners alongside the DWP and FCA. And decumulation, already live given tPR’s parallel work on guided retirement, was elevated to what tPR’s consultation response calls the most significant system-wide challenge the strategy addresses. The AI governance link was strengthened too, tying the strategy’s commitments to technological development sitting within a context of appropriate safeguards and accountability. The substantive AI expectations were largely already public, though, set out in a standalone announcement two months earlier.
The AI Governance Expectations, in Detail
tPR’s most concrete AI position doesn’t sit in the strategy document at all. It came in a press release on 20 May 2026, alongside a document-library “AI plan” setting out four founding principles: outcome-focused, meaning AI use has to be safe and serve members’ interests; principles-based, judged against fairness, accountability, transparency and explainability rather than a fixed technical rulebook; technology-agnostic and evidence-based, so tPR monitors actual adoption and risk rather than naming specific tools; and ecosystem collaboration with government, other regulators and industry as practice develops.
From those principles come six expectations for trustees and scheme managers: clear governance structures for AI use, extending to administrators and other providers acting on a scheme’s behalf; thorough testing and ongoing evaluation of any AI system, before and after deployment; risk management that identifies threats and reviews protective measures regularly; vigilance against AI-enabled scams, with a swift response as new techniques emerge; a clear data strategy built on high-quality data; and appropriate professional advice when evaluating or deploying a solution.
The point most worth flagging to clients is accountability. tPR’s position, echoed since in industry commentary, is that trustees remain responsible for outcomes even where an administrator is the one running the AI system. Delegating the task doesn’t delegate the accountability for it. Trustees aren’t expected to become technology experts, but they are expected to explain, in governance terms, what an AI system is doing on their scheme’s behalf and why they’re confident in it. tPR’s chief executive, Nausicaa Delfas, put the logic plainly: “Trust is the most valuable asset in our system, and that trust depends on the safe and responsible adoption of AI in members’ interests.”
None of this is abstract to tPR itself. Its own AI-enabled scam detection work has assessed more than 2,000 websites and contributed to the removal of 29 high-risk sites, cutting manual scanning and triage by around two hours a day. It is a defensive tool for the regulator, and a capability scammers are turning against members too. Detailed AI adoption guidance is due later in 2026, following an industry consultation over the summer; what exists today is principles and expectations, not yet a finished rulebook.
What It Means for Advisers
Keep two audiences straight, as tPR’s own framing does. For scheme sponsors weighing up providers, AI governance now belongs on the same due diligence checklist as investment governance and cyber resilience. Ask not just whether a provider uses AI in member communications or scam screening, but who inside its governance structure is accountable for it and how it has been validated. For trustee clients, the accountability point is the one to land hardest. They cannot outsource responsibility for an AI system by outsourcing the system itself, so any provider contract should set out how oversight and explainability actually work rather than simply asserting that safeguards exist.
Because tPR’s detailed guidance is still to come, this is a good moment to get ahead of it with clients already using AI in administration or member-facing tools, rather than waiting for a finished standard to react to. And because the strategy ties AI governance maturity to the same fewer-larger-better-run direction driving its wider value for money agenda, expect it to become one more factor in whether a smaller scheme can keep pace with tPR’s expectations, or whether a move to a larger, better-resourced platform is the more defensible long-term position.
Things to reflect on for CPD
- When you next review a corporate client’s scheme governance, how would you check whether the trustees or the administrator can show a documented AI governance framework, rather than a general assurance that AI is being used carefully?
- Which of your scheme clients have already deployed AI in member communications, scam detection or administration, and do you know who inside their governance structure has been made accountable for it?
- How would you factor tPR’s consolidation and value for money direction into a conversation with an employer client weighing whether to stay with a smaller scheme or move to a larger, more heavily governed one?
- What’s your process for keeping clients briefed as tPR’s detailed AI adoption guidance, due later in 2026, moves from stated expectation to published standard?
