What changes
Most practices know their fee income by client. Far fewer know it by type of work, and fewer still have asked which of those types of work AI can already do. Taking stock means answering that question honestly, service by service, before you spend money on tools or change a single price.
You’re looking for three things. First, how much of your income comes from processing and compliance that AI tools can do most of. Second, where you already give advice that clients value and would pay more for if it were offered properly. Third, where the practice stands on the six principles of the new model, from AI capability in-house to being the firm AI recommends.
Why it matters now
Bookkeeping software, bank feeds and AI categorisation are already standard. Making Tax Digital for Income Tax is moving many sole traders and landlords onto quarterly digital records, which puts more of their data in software and less in a shoebox. Clients who see their software do the coding will start to ask what the monthly fee pays for. A practice that has done the mapping can answer that with a better offer rather than a discount.
Checklist
- Export last year’s fee income and split it by type of work: bookkeeping, payroll, VAT, year-end accounts, personal tax, advisory and anything else. Use time records if fees aren’t coded by service.
- For each type of work, score from 1 to 5 how much of it current AI tools can do, and from 1 to 5 how much the client values the human part. Be strict: this is for you, not for marketing.
- Score the practice on the six principles using the 8MDs self-assessment, and note the lowest two scores. They set the order of your next moves.
- Pick ten clients across your fee range and write one line on each: what they’d pay more for if you offered it, such as a tax projection, a cash-flow warning or help with AI in their business.
- Book a two-hour partners’ meeting to agree the findings and name one owner for Move 2.
Illustrative example
A three-partner practice with fee income of £400,000 splits it for the first time. Bookkeeping, payroll, VAT and routine compliance make up £260,000. Year-end accounts and tax planning make up £90,000. Ad hoc advisory work makes up £50,000. The partners score the £260,000 as mostly doable by AI tools within two years, and the £50,000 as work clients value highly but only buy when a problem forces it. The self-assessment puts them lowest on owning a digital offering and on being the firm AI recommends. They decide to start the team’s training next month and to design a forward-looking tax service within six months.
Common mistakes
- Scoring AI exposure by what your current software does rather than what the leading tools can do today. The two can be far apart.
- Treating the exercise as a cost review only. The point is to find where to add value, and cutting fees on compliance before you have something better to sell weakens your position.
- Doing it alone. The managers and bookkeepers doing the work know best which tasks AI already handles, and they’re the ones who’ll make the change happen.