What changes
In most firms AI use is patchy. One or two people use a chatbot to draft emails, others avoid it, and nobody is quite sure what client data is allowed near it. Building capability means making AI a skill the whole firm has, in the same way everyone can use the practice management system. The first principle of the New Model is that the firm understands AI properly, uses it every day and can deploy it, rather than buying it as a black box it can’t question.
In practice that means training for everyone, written rules for safe use and a named person who owns the firm’s progress. Each depends on the others: training without rules leads to risky habits, and rules without training are ignored.
Why it matters now
Every later move depends on this one. You can’t hand over the legwork safely if staff don’t know how to check AI output, and you can’t build your own client tools if nobody in the firm understands what they’re commissioning. Capability also protects you. Whatever tool helped produce a piece of work, the FCA, the SRA, ICAEW and ACCA all hold the firm responsible for the advice and documents it puts its name to. A trained team working to clear rules is the best evidence that you’re using AI with care.
There’s a practical reason too. The tools improve every few months. A team that understands the fundamentals can pick up each improvement quickly, while a team that only learnt one product’s buttons has to start again each time.
Checklist
- Enrol everyone who touches client work. Start with the fundamentals module for your profession, then the module closest to each person’s role. Set a completion date no more than eight weeks away.
- Write a two-page AI use policy. Cover approved tools, what client data may and may not be entered, how output is checked and recorded, and who to ask when unsure. Check it against your data protection obligations and your regulator’s current guidance.
- Pay for business accounts on approved tools. Choose plans where your data isn’t used to train the provider’s models, and ask staff to stop using personal accounts for work.
- Name an AI lead with protected time. Two hours a week is enough to start. The lead collects what works, keeps the policy up to date and runs a short monthly session where colleagues show each other what they’ve tried.
- Start a shared prompt library. Save prompts that work well on your own tasks, each with a note of what the output must be checked for before it’s used.
Illustrative example
A ten-person law firm runs its capability programme over eight weeks. Each person spends an hour a week on training, which is 80 staff hours in total. The AI lead, a senior associate, spends a further two hours a week, adding 16 hours, so the programme takes 96 hours of the firm’s time. Business licences for an approved assistant cost £25 a seat per month, which is £250 a month or £3,000 a year for the whole firm.
By week eight every fee earner has used AI on at least five real tasks under the policy, the prompt library holds 30 tested prompts for the firm’s most common letters and research questions, and the firm has a shortlist of the processes it will hand over in Move 3.
Common mistakes
- Training only the keen. If half the team opts out, processes can’t change. Make the fundamentals module part of everyone’s role, including partners and support staff.
- Banning AI instead of setting rules. A ban usually pushes people onto personal accounts where you can’t see what’s happening. Clear rules and approved tools are safer.
- Treating the policy as finished. New tools and new regulatory guidance will keep arriving. Review the policy every quarter and make the AI lead responsible for the updates.