There are two easy stories about AI and jobs.
The first says everything will be fine. Technology makes work cheaper, the economy grows, people move into better roles, and the disruption eventually settles into progress.
The second says the opposite. AI will wipe out millions of jobs, hollow out professions, and leave people stranded before society has even understood what is happening.
Both stories are too tidy.
The more uncomfortable truth is that AI does not affect every occupation in the same way. Some markets expand when work becomes cheaper. Some contract. Some workers are complemented by technology. Others are substituted by it. Some businesses use AI to create new value. Others use it mainly to reduce headcount, protect margins, and do the same work with fewer people.
The pie may get bigger.
But the pieces do not redistribute on their own.
That matters far beyond the labour market. It matters in financial advice, too.
Because in financial advice, AI is already changing who does the work, who bears the cost, and who captures the benefit.
Source: The pie gets bigger. The pieces don’t redistribute on their own. Jon Twigge · 8 June 2026.
The financial advice version of the AI question
The usual AI and jobs question is:
“How many jobs will AI take?”
But that is not the most useful question.
A better question is:
“When AI makes a service cheaper to produce, who gets the benefit?”
Does the client pay less?
Does the adviser work less?
Does the firm preserve or increase its margin?
Or does the whole industry continue charging as if nothing material has changed?
That is the question now facing the financial advice sector.
For decades, much of the advice industry has been built around a simple commercial model: gather assets, place them into products, and charge an ongoing percentage fee for continued advice, review, access, reassurance, or relationship management.
The language is “advice”.
But in many cases, what is being sold is not advice in the purest sense of the word.
Pure advice would be independent judgement, paid for directly, with no structural requirement that the client buys a product, transfers assets, or stays inside a controlled proposition. Pure advice would include the possibility of saying:
“Do nothing.”
“Keep what you have.”
“Pay down debt.”
“Use cash.”
“Run your own planning.”
“You do not need an ongoing percentage fee.”
Much of the market is not designed around that freedom. It is advice-shaped distribution. It is product intermediation wrapped in relationship language. It is a pathway through which consumer assets are gathered, retained, and charged.
That distinction matters more now because AI is eroding the old justification for many advice fees.
The work has changed. The price has not.
Clients are already arriving at meetings better prepared.
They have used AI to organise their questions. They have modelled scenarios. They have compared options. They have read explanations. They have sense-checked jargon. They have asked the questions they were once too embarrassed to ask.
The old information asymmetry is weakening.
That does not mean human judgement has no value. It does not mean professional expertise has disappeared. It does not mean people should blindly trust AI with complex financial decisions.
But it does mean the client’s capability has changed.
And if the client’s capability has changed, the value equation has changed.
The uncomfortable question is this:
If AI now does a large part of the preparation, explanation, modelling, analysis, and client education that previously justified the adviser’s fee, why has the fee not changed?
If the adviser can serve more clients in fewer hours, who captures that efficiency?
If the work becomes cheaper to produce, why should the client continue paying as though it has not?
This is where the wider AI labour-market debate becomes directly relevant. AI may grow the economic pie. But without intentional redistribution, the benefit flows to whoever controls the commercial model.
In financial advice, that is usually not the client.
St James’s Place and the power of the relationship machine
St James’s Place is the most obvious case study because it is not a marginal player. On its own published figures, it manages around 13% of all advised invested wealth and serves around 20% of UK adults who receive financial advice.
The precise figure is not the point.
The point is scale, influence, and systemic importance.
SJP is not just another regulated firm. It is a market leader in advice-led wealth management. Its success, resilience, and client loyalty reveal something important about the deeper psychology of the advice market.
Even after sustained criticism of charges, servicing, performance, exit penalties, and value for money, many SJP clients remain fiercely loyal. Some are not merely loyal. They are defensive. Criticism of the firm can be received almost as criticism of the client’s own judgement.
That is not unusual in high-trust relationship markets.
People do not buy financial advice only as a technical service. They often buy emotional safety.
They buy reassurance.
They buy confidence.
They buy permission not to look too closely.
They buy the feeling that someone else is holding the complexity.
Once that emotional bond is formed, the commercial arrangement can become difficult to challenge. A client may defend the adviser not because the fee structure is objectively fair, but because accepting the opposite would require a painful admission:
“I may have paid too much.”
“I may not have received what I thought I was receiving.”
“I may have outsourced judgement to a system that benefited from my dependency.”
That is hard for anyone to face.
So the defensive response is understandable. It is human. But it is also commercially useful to the industry.
The faith does not fade just because the facts change
This is why the advice market can behave almost like a belief system.
Evidence alone does not necessarily break the spell.
A client can be shown that charges are high. They can be shown that performance has been underwhelming. They can be shown that annual reviews were not properly evidenced. They can be shown that cheaper alternatives exist. They can even be shown that the service they are paying for has changed materially because AI now does much of the heavy lifting.
And still the faith may remain.
Because the relationship is not primarily factual. It is emotional. It is social. It is identity-protective.
The client is not simply defending a wealth manager. They are defending a story about themselves:
“I made a good decision.”
“I chose a trustworthy person.”
“I am sensible with money.”
“I am looked after.”
This is why the old model may continue far longer than critics expect.
Not because it is the best model.
Not because it is the fairest model.
Not because it represents good value in an AI-enabled world.
But because trust, once converted into dependency, is commercially durable.
The regulator’s structural dependency problem
This is also why expecting transformation to come from regulation may be naïve.
The Financial Conduct Authority is funded by the firms it regulates. In the adviser and intermediary fee-block, periodic fees are linked to firms’ annual income. That means the regulator’s own funding ecology is structurally connected to the commercial scale of the regulated market.
This does not prove corruption.
It does not require a conspiracy.
It is simply a structural dependency.
When a market leader is systemically important, the regulator is unlikely to seek a disorderly confrontation with the underlying commercial model. The pressure is more likely to come through disclosure, remediation, record-keeping, Consumer Duty language, fee restructuring, and managed reform.
That is not nothing.
But it is not the same as challenging the deeper question:
Is an advice-led asset-gathering model still fair when AI has shifted capability back towards the consumer?
If SJP or any other large advice-led wealth manager were shown to have built scale on a model that was structurally poor value for many clients, that would not only be a firm-level problem. It would raise questions about the supervisory model that permitted, normalised, and legitimised it.
That is why reform from within the existing system is likely to be slow.
The system has too much invested in its own continuity.
The AI asymmetry
There is another unfairness emerging.
The financial services industry is embracing AI for itself while warning consumers to be careful about using AI for themselves.
Firms will use AI to reduce costs, automate workflows, draft communications, improve segmentation, analyse data, support compliance, generate reports, and increase productivity.
But consumers using AI to understand their own money are often told that AI is risky, cold, impersonal, unregulated, and dangerous.
There is truth in the warning. AI can be wrong. It can hallucinate. It can oversimplify. It can miss context. It should not be treated as an infallible adviser.
But the warning becomes self-serving when it protects professional dependency rather than consumer safety.
The message becomes:
“AI is safe when we use it to manage you. It is unsafe when you use it to question us.”
That is capability asymmetry.
And restoring human agency means challenging it.
Consumers should not be encouraged to hand blind trust to AI. But neither should they be discouraged from using AI to become more capable, more prepared, more questioning, and less dependent.
The future is not adviser versus AI.
The future is agency versus dependency.
What happens when the spell fades?
The spell will not fade all at once.
It will fade slowly, unevenly, and probably first among the people who are already asking better questions.
Not:
“Do I like my adviser?”
But:
“What am I actually paying for now?”
Not:
“Is my adviser a good person?”
But:
“Is this commercial arrangement still fair?”
Not:
“Can AI replace advice?”
But:
“Can AI help me understand enough to stop surrendering control?”
This is the shift the industry underestimates.
Clients do not need to become investment experts. They do not need to build portfolios from scratch. They do not need to reject human support.
They simply need enough capability to see the deal clearly.
Once that happens, the old percentage-based model starts to look very different.
A recurring fee deducted from life savings is no longer seen as normal simply because it has always been normal. It becomes a claim that must be justified.
What service was delivered?
What value was created?
What alternatives were considered?
What work did the adviser do?
What work did the client now do themselves?
What work did AI do?
And who captured the saving?
Advice out. Agency in.
The Academy of Life Planning exists to restore human agency in financial and life decisions.
That does not mean replacing one dependency with another. It does not mean telling people to trust AI instead of advisers. It does not mean attacking every adviser or pretending that all professional relationships are exploitative.
It means something simpler and more radical.
People should become progressively less dependent on professional authority, not more.
Good expertise should make itself less necessary over time.
Good planning should leave the client clearer, stronger, and more capable.
A fair model should allow people to pay for help when they need it, without surrendering an indefinite percentage of their wealth simply to remain inside a relationship.
AI makes that possible.
Not because AI is perfect.
But because AI can help close the capability gap. It can help people prepare, understand, question, organise, model, and reflect. It can help them become active participants in their own planning rather than passive recipients of professional reassurance.
That is the real redistribution question.
When AI makes financial planning cheaper, faster, and more accessible, will the benefit be captured by advice firms through higher margins?
Or will it be returned to consumers through lower dependency, clearer pricing, and greater agency?
The pie may get bigger.
But unless we choose differently, the pieces will not redistribute on their own.
Academy of Life Planning 