“The question is no longer whether to allow AI, it is who is AI going to serve?… the answer must be citizens.”
— Sheldon Mills, AI and the Future of Retail Financial Services (2026)
The publication of the Mills Review marks an important moment in the evolution of financial services.
For the first time, a major regulatory review acknowledges something many of us have been observing for several years: AI is not simply going to transform financial firms. It is going to transform the relationship between financial firms and the people they serve.
The Review concludes with seven priority recommendations. The final recommendation may prove to be the most important:
Develop a trusted public-interest AI-enabled financial capability service.
We believe this recommendation deserves far more attention than it has received.
The report explains why
Throughout the Review, Sheldon Mills builds a compelling case for why citizens need AI on their side.
He points to longstanding failures in retail financial services:
- only 9% of consumers use traditional financial advice
- significant protection gaps
- low switching rates
- financial exclusion
- billions of pounds sitting in unsuitable savings accounts.
The Review argues AI could help people make better financial decisions because it can combine an individual’s own information with wider knowledge to provide personalised insight and guidance.
Consumer research found people will only adopt these technologies if three conditions are met:
- trust
- control
- access.
The report also recognises something profound.
As AI develops, consumers will increasingly delegate routine financial tasks to AI agents acting on their behalf. But they must still be able to understand, oversee and challenge those systems.
These are exactly the right questions.
But one question remains unanswered
The report explains why a public-interest AI-enabled financial capability service is needed.
It says much less about what such a service actually does.
If we are to build capability rather than dependency, what capabilities should citizens actually receive?
A practical interpretation
For the past fifteen years, the Academy of Life Planning has been developing what has become Academy OS.
Not because we anticipated the Mills Review.
But because we recognised the same problem from a different direction.
Financial institutions have invested billions helping organisations make better decisions.
Citizens have received calculators, comparison sites and generic guidance.
There has never been equivalent investment in helping individuals think more clearly about their own lives.
Academy OS was designed to begin closing that gap.
Rather than replacing human judgement, it helps people strengthen it.
In practical terms, that means helping people:
- understand complex financial documents in plain English
- question products, contracts and firms before signing
- investigate investments and identify risks before committing money
- organise their thoughts before making important decisions
- prepare for conversations with advisers, providers or regulators
- remain the decision-maker, with AI acting as a capability amplifier rather than a substitute.
This is not advice.
It is not product recommendation.
It is not discretionary decision-making.
It is capability.
From the advice gap to the agency gap
The Mills Review describes the opportunity to close the advice gap.
We would argue the opportunity is even bigger.
The deeper issue is an agency gap.
For decades, financial services have largely been organised around information asymmetry.
Professionals knew more than consumers.
Institutions possessed more data than citizens.
Technology was concentrated inside firms.
AI changes that.
For the first time, ordinary people can have access to sophisticated tools that help them understand complexity before making decisions.
The challenge is ensuring those tools are genuinely working for the individual, not simply becoming another channel through which institutions influence behaviour.
AI should increase capability, not dependence
One section of the Review describes how people’s roles will change as AI becomes more autonomous.
Instead of making every decision themselves, people will increasingly set boundaries, grant permissions and oversee outcomes.
That is an important observation.
But it also raises a new question.
How do people know what boundaries to set?
How do they know what permissions to grant?
How do they recognise when outcomes should be challenged?
Those are not technical questions.
They are capability questions.
No AI agent can restore agency unless the individual understands enough to remain meaningfully in control.
Building AI that serves citizens
Perhaps the most memorable sentence in the Review comes in Sheldon Mills’ foreword:
“The question is no longer whether to allow AI, it is who is AI going to serve?… the answer must be citizens.”
We agree.
The Academy of Life Planning has never sought to build AI that replaces advisers or competes with firms.
Our aim has been different.
To build AI that helps citizens become more capable.
To help people ask better questions.
To understand before they decide.
To arrive at conversations better prepared.
To remain firmly in control of their own lives.
The Mills Review has identified an important public policy destination.
Academy OS is not the public-interest AI-enabled financial capability service the Review envisages.
But we believe it demonstrates one practical interpretation of what that future could look like.
Because restoring human agency has never been more important.
And if AI is going to serve citizens, it should begin by helping them think—not by thinking for them.
Academy of Life Planning 