By Steve Conley, Founder, Academy of Life Planning | 12 October 2026
Financial planning firms are becoming better at using AI. They can capture conversations, organise information and produce documents with less manual work. That is a real improvement.
But a faster process does not tell us whether the person at its centre understands their situation any better.
This week’s developments expose that distinction. The FCA is examining agentic AI in firms. New international research shows clients bringing AI into financial conversations. Investment is flowing into technology that connects the firm’s systems and client data.
Together, they raise a question for every planner: when your technology improves, whose capability grows?
Institutional AI completes processes for the firm. Sovereign AI builds capability for the person.
At the Academy of Life Planning, we use that distinction to describe two design priorities. Institutional AI serves an organisation’s workflow. Sovereign AI helps a person understand, shape and retain control of their own planning. Neither outcome follows automatically from installing an AI tool.
The regulator is looking at what agents do
At the FCA’s Annual Public Meeting on 6 October, the regulator described an AI lab cohort focused on agentic technologies: systems that can carry out tasks rather than simply answer questions. It also said it had visited firms to examine deployment and expected to publish good and poor practice examples shortly.
This is a confirmed statement about supervisory work, rather than a new rule.
The FCA also reiterated that consumer use of general-purpose large language models sits outside its perimeter. That statement should be read in context. It does not determine the status of every purpose-built financial tool or activity.
Source: FCA Annual Public Meeting transcript, published 9 October 2026
For planners, the practical issue is what happens between an AI output and a person’s action. Can they see its assumptions? Can they challenge it? Do they know when to seek professional help?
An agent that completes a task efficiently can still leave its user poorly equipped to judge the result.
Clients have AI too
FPSB’s global research, released on 7 October, surveyed 8,095 financial planners across 37 territories. Thirty-three per cent reported clients using AI for general financial education, while 25% reported its use to validate or compare professional advice.
These are planners’ reports, not a direct survey of client capability. Nevertheless, they show that AI is entering the conversation from both sides.
The study also found that 93% of planners said AI had improved their work efficiency. The profession is experiencing the benefits while encountering clients who arrive with their own analysis.
Source: FPSB’s 2026 global research
A client bringing an AI-generated report is bringing something to examine together. The planner can ask where the information came from, which assumptions drive the conclusion and what remains uncertain.
That conversation can strengthen understanding. It can also reveal an error before it becomes an expensive decision.
The opportunity for a Total Wealth Planner is to help the person become better at this process over time. Professional judgement remains valuable where complexity, consequence or uncertainty calls for it.
Investment follows the firm’s workflow
On 7 October, Mercia announced a £4.3 million funding round for PlannerPal. Its description emphasises connecting conversations, documents and adviser systems into a client intelligence layer. Mercia reports use by more than 250 UK advice firms; that reach is an investor announcement claim.
Source: Mercia’s funding announcement
There is a clear commercial logic here. Firms have fragmented information and administrative costs. Technology that addresses those problems has an identifiable buyer and a measurable return.
My interpretation is that this also reveals the challenge for client-owned planning: its value needs equally concrete evidence.
Can the person explain their plan? Change an assumption? Compare possibilities? Carry their information forward when a professional relationship ends?
Those are useful measures of capability alongside time saved.
Defaults need to be understood
In an 8 October speech, The Pensions Regulator urged trustees to prepare for guided retirement and the design of default pension paths around member needs. It said a joint DWP/FCA consultation would follow this autumn. The detailed framework remains forthcoming.
Well-designed defaults can help people who would otherwise struggle to choose. A Total Wealth Plan can help someone understand how a default fits their circumstances, other resources and priorities. Specialist advice may be needed to assess consequential pension choices.
A practical test for the planning model
For Academy OS and Total Wealth Plans, the design ambition is continuous agency with episodic expertise: people can work with their own information and understanding between occasions when professional support is valuable.
That ambition must be tested in practice. A useful starting point is to ask a client to explain one important assumption, model a change and identify a question that needs further help.
If they cannot, the next task is to improve their understanding.
This week’s announcements do not establish a new professional standard or prove that AI creates agency. They show why building agency needs deliberate attention.
The planner of the future will not be the client’s operating system. They will help the client operate one.
Does your AI make you faster—or your clients more capable?
Book a 50-minute Zoom call with Steve Conley for £50 to explore what agency-led planning could look like in your practice, where your human judgement adds value, and one practical change you could test.
