What changes
In many firms, AI arrives through one or two enthusiasts using a free chatbot on their own. That gives the firm the risk without the benefit. Building capability means everyone, from the principal to the newest administrator, understands what these tools do well, where they get things wrong and how to check them. It also means the firm decides which tools are approved, what client data can go into them and who signs off the output.
The aim is a firm that can deploy AI itself rather than buying it as a black box. You don’t need developers for that. You need people who can write a clear instruction, spot a confident but wrong answer, and turn a task that works once into a process that works every time.
Why it matters now
Every later move depends on this one. You can’t hand the legwork to AI, build your own client tools or reprice with confidence if the team doesn’t trust the output or know how to test it. AI bought as a black box also leaves the firm dependent on a supplier’s choices about data, cost and quality, with no one inside who can judge them.
Your obligations don’t change when a tool does the work. The firm stays responsible for the suitability of its advice and for handling client data properly under UK GDPR. Under the Consumer Duty, communications still have to support client understanding, whoever drafted them. A written AI policy and a trained team are how you show those responsibilities are being met.
Checklist
- Enrol the whole team on the 8MDs adviser course, starting with AI Fundamentals for Financial Advisers and Practical AI Use, and book an hour a week in everyone’s diary to work through it.
- Write a one-page AI policy covering approved tools, what client data may be used, how outputs are checked and who is accountable for each use.
- Choose a business-grade AI tool whose terms confirm your data isn’t used to train its models, and stop the use of personal accounts for client work.
- Run a weekly 30-minute session where one person shows a real task they have done with AI, including what went wrong and how they caught it.
- Name an AI lead who keeps the policy current, tracks which tasks have moved to AI and reports progress to the directors each quarter.
Illustrative example
A firm of six people, made up of two advisers, two paraplanners and two administrators, sets aside one hour a week each for eight weeks. That’s 48 hours of team time in total. The first four weeks cover the course modules. In the last four, each person picks one task from the Move 1 list and writes a tested prompt and a checking routine for it.
By week eight the firm has six tested routines, including a meeting summary, a first-draft review letter and a provider chase email. Both advisers sign off the AI policy, and the paraplanners check each routine’s output against a short written list before anything reaches a client or a client file.
Common mistakes
- Training only the advisers. Much of the time saving sits with paraplanners and administrators, so they need the skills as early as anyone in the firm.
- Banning AI until the policy is perfect. Staff may use it anyway on personal accounts, where you can’t see or control it. A short policy now is worth more than a long one next year.
- Treating a good demo as a finished process. A task isn’t ready to hand over until it works on real files and has a written checking step that someone owns.