On 25 August 2026, McKinsey published two complementary pieces on the next stage of artificial intelligence: its global survey, The State of AI in 2026: On the Road to ROI, and a shorter commentary, Four Perspectives on What Matters Now in AI. Together, they reveal an important tension. AI adoption is accelerating, individual productivity and decision-making are improving, and organisations are beginning to redesign workflows and build more of their own technology. Yet enterprise-level financial returns remain much less widespread. McKinsey’s findings suggest that the real breakthrough may not come from simply inserting AI into existing processes, but from redesigning how work is done around what AI now makes possible.
The First Dividend From AI May Be an Agency Dividend Before It Becomes a Financial Dividend
Most organisations are asking the same question about AI:
Where is the return on investment?
It is an understandable question.
Billions are being invested. New tools are appearing every week. Employees are using AI to write, analyse, code, research and automate.
Yet the financial returns have not arrived at the same speed as adoption.
McKinsey’s latest global research captures the gap clearly.
Around 80% of respondents say AI has improved their individual productivity. Around half say it helps them make better decisions.
But only 37% say AI has made a positive contribution to their organisation’s earnings.
Perhaps we are looking for the wrong dividend first.
The first dividend from AI may not be a financial dividend.
It may be an agency dividend.
Productivity is not the same as agency
Much of the AI conversation is still framed around productivity.
Can the employee write the report faster?
Can the adviser produce the recommendation faster?
Can the company serve more customers with fewer people?
Those things matter.
But there is another possibility.
What happens when AI helps a person understand something they previously could not understand?
What happens when somebody can investigate their own problem?
Model their own future?
Challenge an assumption?
Explore alternatives?
Prepare themselves for a professional conversation?
Or make a decision without automatically having to delegate that decision to somebody else?
That is not merely productivity.
It is an increase in the person’s capacity to act.
That is what we mean at the Academy of Life Planning by human agency.
Understanding creates options.
Options create choice.
Choice creates action.
And action creates agency.
McKinsey is finding the beginnings of this
The numbers are revealing.
While enterprise financial returns remain relatively concentrated, people themselves are already reporting substantial benefits from AI.
Eight in ten respondents say it has increased their productivity.
Around half say it has helped them develop skills and make better decisions.
That is significant.
Before AI transforms the balance sheet, it may be transforming the individual.
This matters particularly in financial services.
For decades, access to sophisticated financial knowledge has largely depended on access to an expert.
The individual has information.
The professional has expertise.
The individual therefore gives the professional their information and waits for an answer.
AI begins to change that relationship.
It does not eliminate the value of expertise.
It changes where expertise is required.
From continuous expertise to continuous agency
The traditional financial advice model implicitly assumes that expertise needs to sit continuously between the person and many of their financial decisions.
The resulting architecture often looks something like this:
Client → adviser → institution → product
Information flows upwards.
Recommendations flow downwards.
The professional interprets.
The client receives.
AI makes another architecture increasingly possible:
continuous agency, episodic expertise.
The individual can increasingly maintain their own information, ask questions, explore possibilities, run calculations, test assumptions and prepare decisions continuously.
Human expertise can then enter where it adds most value:
when circumstances become complicated,
when trade-offs are difficult,
when judgement matters,
when emotions are interfering with decisions,
when regulation or specialist knowledge is required,
or when the consequences of getting something wrong become significant.
That is a very different relationship.
The professional does not disappear.
The professional becomes more valuable precisely because their time is no longer consumed doing work technology can increasingly help the person do themselves.
Don’t put AI into the old advice machine
One of McKinsey’s most important findings concerns organisations that are actually generating significant value from AI.
The organisations they describe as AI high performers are not simply using more AI.
They are redesigning workflows around what AI now makes possible.
They are 3.3 times more likely than others to intend to use AI to fundamentally transform their organisation.
Nearly three-quarters report fundamentally redesigning workflows because of AI, compared with around a quarter of other organisations.
That distinction matters enormously.
Putting an AI assistant beside a financial adviser may make the adviser faster.
AI can draft correspondence.
Summarise meetings.
Produce research.
Complete paperwork.
Prepare reports.
Those are useful efficiencies.
But they leave the architecture largely unchanged.
The adviser remains the centre of the system.
The more interesting question is:
What would financial planning look like if we redesigned the workflow around what AI now enables the individual to do?
That is where transformation begins.
The client can become part of the operating system
Imagine financial planning beginning not with an adviser gathering information from a client, but with the person maintaining their own permanent life record.
Their goals.
Values.
Finances.
Family.
Work.
Health.
Aspirations.
Concerns.
Decisions.
Their data remains theirs.
AI helps them understand it.
Specialist applications help them explore particular decisions.
The planner enters the process when human expertise is valuable.
And every interaction leaves the individual more capable than before.
That is very different from traditional financial services, where information is normally gathered into the operating system of the institution.
The Academy model reverses this.
The operating system sits around the person.
We call it Academy OS.
Its role is not simply to make financial planners more efficient.
Its purpose is to help people become more capable.
Small practices can now build what they previously had to buy
There is another important finding buried within the McKinsey research.
Nearly one-third of organisations surveyed say they have already decided not to purchase at least one software product or feature because they could build the functionality themselves using agentic coding tools.
This could have profound consequences for professional services.
Historically, a small financial planning firm had little meaningful control over its technology.
It bought a CRM.
Bought a cashflow modelling package.
Bought risk profiling software.
Bought research.
Bought platform technology.
Bought compliance systems.
Bought whatever the vendors serving the industry decided to manufacture.
That creates dependency.
AI changes the economics of software creation.
A small practice can increasingly create its own:
tools,
workflows,
calculators,
decision-support systems,
knowledge bases,
client interfaces,
and intellectual property.
It does not mean every practice should build everything.
McKinsey makes that point too.
The strategic question becomes:
What should we buy, what should we build, and what capability do we need to own?
At the Academy of Life Planning, we have already started answering that question.
We buy commodity infrastructure.
But we increasingly build the capabilities that define our proposition.
That is what Academy OS represents.
Buy the infrastructure. Own the intelligence.
For decades, the competitive advantage of small professional firms often came from relationships and expertise.
Technology belonged to large institutions.
AI changes that balance.
A small professional practice can now encode its methods.
Turn frameworks into applications.
Turn experience into workflows.
Turn intellectual property into interactive tools.
And continuously improve those tools without requiring an enterprise software budget.
That gives us a useful principle:
Buy the infrastructure. Own the intelligence.
The generic components can come from elsewhere.
Cloud computing.
AI models.
Storage.
Communications.
Payment processing.
But the distinctive reasoning, methods, frameworks and client experience can increasingly belong to the practice itself.
That is potentially one of the great levelling effects of AI.
Academy OS is our answer
This is why we have been building Academy OS.
It is not simply a collection of AI tools.
It is an attempt to build a different architecture for financial and life planning.
At its centre is the individual.
Their information.
Their goals.
Their choices.
Their agency.
Around them sit different layers of support.
Their personal life record.
Our GAME Plan life-planning framework.
AI-assisted exploration.
Specialist applications.
Human Total Wealth Planners.
And specialist expertise when required.
The system might be represented simply as:
My Life Record → GAME Plan → AI exploration → specialist tools → human expertise → action → updated Life Record
The important thing is where the intelligence accumulates.
It accumulates around the individual.
Not solely around the institution.
This is more than a technology strategy
McKinsey makes another important observation.
The best-performing organisations are not succeeding because they have accumulated the most AI tools.
Their advantage comes from the coherence of their approach.
They combine workflow redesign, leadership, human oversight, measurement and risk management.
McKinsey argues that the limiting factor is increasingly becoming an organisation’s ability to absorb change rather than the technology itself.
That should make every professional services firm think carefully.
The question is no longer:
Which AI tools should we buy?
It is:
What kind of organisation should we become now these capabilities exist?
And for financial planning:
What kind of relationship between person and professional should now exist?
AI may not remove humans. It may change where humans matter.
There is also a useful warning in McKinsey’s research.
Predictions about AI-driven workforce reductions have so far exceeded reality.
Last year, 32% of respondents expected AI-related workforce reductions.
Only 14% subsequently reported that AI had contributed to an overall decline in workforce size.
That does not mean employment will remain unchanged.
It suggests the transformation may be subtler than simply replacing people with machines.
Human work may move.
Routine cognition can increasingly move towards AI.
Human contribution can increasingly move towards:
judgement,
wisdom,
interpretation,
challenge,
empathy,
reassurance,
creativity,
accountability,
and navigating ambiguity.
This is particularly relevant to financial planning.
The future Total Wealth Planner may spend considerably less time producing information.
And considerably more time helping people decide what that information means for their lives.
The real productivity breakthrough may belong to the client
There is an assumption hidden inside much of the financial services AI conversation.
The objective is to make the adviser more productive.
One adviser can serve more clients.
One institution can process more customers.
One call centre can answer more questions.
But why should all of the productivity dividend accrue to the provider?
What happens if some of it accrues to the individual?
If a person becomes capable of doing things for themselves that previously required professional intervention, something important happens.
The cost of support can fall.
The dependency on professionals can fall.
Access can expand.
Expertise can be reserved for situations where expertise genuinely matters.
And the individual becomes stronger.
That is the agency dividend.
Perhaps we are measuring AI too narrowly
Organisations naturally want to know whether AI improves revenue or reduces costs.
They should.
But perhaps there is another metric worth watching.
Not simply:
How much money did AI save us?
But:
How much more capable did AI make the person?
Can they understand more?
Choose more intelligently?
Challenge more confidently?
Solve more problems themselves?
Know when to seek help?
And enter professional relationships as a participant rather than a dependent?
Those outcomes may eventually create financial returns too.
But they matter before that.
The architecture is becoming possible
For many years, a model based on continuous personal agency and episodic professional expertise would have been difficult to deliver.
The technology was too expensive.
Knowledge was difficult to distribute.
Software development required specialist teams.
Professional expertise was difficult to replicate.
Information sat in fragmented institutional systems.
That is changing.
McKinsey’s research shows organisations increasingly building their own tools.
Redesigning workflows.
Using AI to improve individual decision-making.
And learning that simply inserting AI into old processes is not where the greatest value lies.
In other words, the technology and economics are beginning to catch up with a different philosophy of professional support.
One where the person does not permanently outsource their thinking.
One where technology strengthens the individual rather than simply strengthening the institution.
One where humans remain available when human expertise genuinely matters.
Continuous agency. Episodic expertise.
The first dividend from AI may not appear on a company’s balance sheet.
It may appear in a human being who understands more, sees more choices and feels able to act.
And if we design the technology well, that may turn out to be the most valuable dividend of all.
Academy of Life Planning 