Why the Mills Review mattered beyond financial intermediation
By Steve Conley, Founder, Academy of Life Planning | 19 October 2026
A person opens a letter from their pension provider. They understand the words, but cannot see what the assumptions mean for them.
Another receives an adviser’s proposal. They want to examine the charges, alternatives and reasoning before agreeing.
A third has a complaint rejected. They need help organising the evidence and working out which questions remain unanswered.
None of these people necessarily needs someone to sell them an investment. They need the capability to understand, question, choose, act and retain control.
That is why an important part of the Mills Review deserves more attention than today’s headlines about warnings on AI-generated financial advice.
The proposal behind the headline
Published in July, the Mills Review recommended that the FCA consider convening a trusted, free, inclusive AI financial capability service with government, the Money and Pensions Service, consumer bodies and industry. It addressed access to reliable support and the risks of advanced AI being available mainly through paid services or commercially conflicted channels.[1]
Today, Jessica Rusu announced that the FCA would build on existing initiatives instead of creating the proposed tool.
She highlighted work helping investors spot scams, verify information and challenge unreliable AI outputs. The speech does not set out a detailed explanation for choosing that route.[2]
Those activities have value. But their scope leaves a larger question:
Who equips citizens to scrutinise legitimate institutions—not just avoid illegitimate ones?
Legitimacy does not remove the need for scrutiny
A legitimate institution can offer a proposition that deserves searching questions.
What assumptions drive this projection? What does the total charge include? Which alternatives were excluded? What would make this recommendation unsuitable? What evidence supports the rejection of my complaint?
These are ordinary questions of informed participation.
Authorisation does not relieve a customer of the need to understand a transaction. Nor should it relieve an institution of the need to explain one.
Trust becomes more meaningful when the person can examine the basis for it.
Capability is a distinct function
Much of the debate about AI in finance starts with an institutional question: which parts of an adviser’s work can a machine perform?
That is a reasonable question for advice businesses. It does not describe the whole opportunity.
AI can also help a person read a document, identify missing information, compare explanations, test an assumption or prepare for a conversation. It can help them recognise uncertainty and decide when to seek specialist expertise.
The output might be a better question. It might be a clearer record. It might be a decision to pause.
Calling all of this “AI advice” can obscure the purpose of the service. It places a tool for examining a proposition into the same discussion as the institution presenting that proposition.
A tool that helps you scrutinise an intermediary is performing a different function from the intermediary itself.
An AI system can perform different functions at different times. We should judge what it actually does, whose interests it serves and how much control the person retains.
What a disclaimer can—and cannot—explain
People should understand a tool’s limitations and the protections available.
However, generic financial guidance does not become regulated investment advice merely because AI delivers it. The FCA’s own guidance distinguishes generic advice from advice on particular investments and includes financial planning among its examples of generic advice.[3]
A warning about the absence of Financial Ombudsman Service or Financial Services Compensation Scheme protection may correct a mistaken expectation. It does not establish whether a capability tool is useful, accurate or appropriate for its purpose.
For such a tool, the immediate questions include: Can I check the sources? Which facts are missing? What remains uncertain? How is my information handled? Can I challenge the answer?
Clear boundaries matter. So does giving people a practical way to work within them.
The deeper issue is the distribution of capability
Institutions have specialist staff, legal resources, extensive records and established processes. AI can increase their ability to analyse, communicate and act.
What happens if the citizen’s side of the relationship does not develop comparable support?
We risk making an unequal relationship more efficient without making it more balanced.
Mills’s proposal matters because it opens a different policy conversation: access to the means of understanding and scrutiny.
A person should not need substantial investable assets to understand their pension. They should not need to buy a product to organise their financial life. They should not have to accept an institution’s explanation simply because they lack the tools to examine it.
Access to capability is a public-interest question.
This is the purpose of Academy OS
At the Academy of Life Planning, we are building towards a continuing capability service that people can use across their financial and wider lives.
Academy OS brings together tools intended to help people organise information, explore options, prepare questions, document decisions and seek support when needed.
The purpose is restoring human agency: helping people understand, question, choose, act and retain control.
We use the idea of a sovereign OS to describe a system organised around the person’s purposes and authority. Sovereignty must be demonstrated through meaningful control, clear permissions, understandable records and practical choices about data and access.
AI does not automatically create agency. A confident answer can encourage dependence as easily as understanding. Our task is to design and test for greater capability, including the ability to question the tool itself.
The overlap with Mills’s proposal is an alignment of purpose. It is not regulatory endorsement. We must earn trust through evidence of what people can understand and do with the support provided.
A practical invitation
The public needs accessible tools and the confidence to question them. Professionals can help people develop that capability and contribute expertise where it adds value. Institutions can welcome informed scrutiny and make their propositions easier to examine.
For policymakers, the test is whether citizens can obtain reliable, accessible support across everyday financial decisions—not simply recognise a scam.
For the Academy, the next step is practical: build, listen, improve and demonstrate whether people leave better able to act for themselves.
The future of finance will be shaped by who can understand and challenge it.
Join the Academy of Life Planning in restoring human agency.
Your plan. Your OS. Your support.
Sources
[1] FCA, The Mills Review: AI and the future of retail financial services, July 2026, priority recommendation 7, pp. 113–117: https://www.fca.org.uk/publication/corporate/the-mills-review.pdf
[2] Jessica Rusu, Building the future of finance in the age of AI, 19 October 2026: https://www.fca.org.uk/news/speeches/building-future-finance-age-ai
[3] FCA, Helping firms provide more support to customers making investment decisions, and PERG 8.26: https://www.fca.org.uk/firms/helping-firms-provide-more-support-customers-making-investment-decisions
