A free, 2–3 hour audit for the managing partner of an independent UK accountancy practice. You get a written six-page report: three things to ship in the next 30 days, a 12-month roadmap, and a plain section on what I'm not recommending and why.
No deck. No follow-up sales pitch. The report is the deliverable. You decide what happens after that.
If you run an independent UK practice with 5–20 staff, the picture is probably familiar. The compliance work is profitable but heavy. Pricing has lagged inflation. Two or three good juniors have come and gone. Advisory revenue is growing, slowly. The firm is fine. The principal is tired.
And then there's AI. Two years of vendor pitches. ChatGPT open in another tab but rarely used in earnest. Karbon tried and abandoned. Receipt Bank stuck. Other firms saying clever things at conferences. The honest position is somewhere between "we should be doing more with this" and "I don't know what I'm even being sold."
MTD ITSA is live and most firms are still working out what it actually means in practice. It belongs in the roadmap. It isn't the reason to do this audit.
The reason is simpler. You'd like a written, operator's view of what AI can actually do for the firm in the next 12 months — and what it can't — from someone who isn't trying to sell you software.
Designed to be read in twenty minutes, printed, marked up, and forwarded. No slides. No frameworks. No abstract diagrams of "the AI-enabled firm." A working document for an operator running a real business.
The whole audit on one page, written so your spouse or your partner can read it over a coffee and understand what I think.
What kind of work, what kind of clients, where the time goes, where the strain is. The bit most consultants skip. The bit that decides whether the rest of the report is useful.
For each: the problem, the recommendation, rough scope, the tools, an effort estimate, and who in the firm should own it. Built to be small enough to actually do before the next quarter-end.
What to do this quarter, what to hold until after self-assessment, and what the firm should look like twelve months from now. MTD ITSA and the January workload shape the sequencing — not a generic AI strategy.
Every audit I've ever seen tells you to do more. This page tells you what to ignore: tools you've been pitched, ideas that sound clever but don't pay for themselves, projects that would be a waste of partner time. Often the most useful page in the report.
I learn the shape of the firm. You learn whether I'm the right person to be doing this. Four questions, one small piece of homework, and a mutual yes-or-no at the end. No obligation either way.
I sit with you and walk through the firm. How clients come in, how the year-end runs, where the bottlenecks are, what's been tried, what the team actually uses. Conversational, not a questionnaire.
Six pages, written. Delivered as a PDF. I'll talk you through it on a 30-minute call if useful — but the report is designed to stand on its own and travel inside the firm without me.
The audit format breaks down for a few specific kinds of firm. I'd rather say no on the scoping call than write a report that doesn't apply.
You don't know me. You've been pitched at by software vendors for twenty years. You won't buy AI work from someone who sends you a deck and a quote, and you shouldn't.
The audit is the demonstration. If the report is useful, you'll know I'm worth talking to about implementation. If it isn't, you've lost half a working day and gained a written document you can use however you like.
In return for the audit, I ask for two things. Honest feedback on the report — what landed, what missed, what was obvious to you that I treated as insight. And permission to use anonymised observations from across the cohort in future writing. No firm is ever named.
I'm an exited founder. I ran ops at Uber India during the launch years, was Head of Ops at Urban Company, founded and sold a B2B marketplace business (acquired by IDX, 2023), and spent the last cycle as Interim COO at Cloud Cycle through their European expansion.
For the last eighteen months I've been doing the same kind of work for founder-led businesses through Shodhan Advisory — embedding inside companies, auditing how work actually moves, and shipping AI workflows into production. The Practice Audit applies that same lens to independent accountancy firms.
I build the things I recommend. When implementation follows an audit, it's delivered through a vetted India-based technical partner I supervise — you have one accountable point of contact throughout, which is me.
There isn't one. The audit is genuinely free, time-boxed, and produces a written report you can use however you like. A subset of firms ask me to help with implementation afterwards — that's a separate, paid conversation, and most don't. The audit stands on its own.
A 20-minute scoping call, followed by 2–3 hours together for the audit itself, plus one optional 30-minute call to walk through the finished report. Most principals find the audit conversation itself is the most useful part — it's the first time they've had a structured discussion about AI with someone who isn't selling them software.
Three reasons. Accountancy workflows — client onboarding, document collection, year-end, payroll — are well-documented and replicable, so what works in one firm tends to travel. Independent practices talk to each other; one good report ripples. And the current AI moment is genuinely confusing for principals — most have been pitched at for two years and bought software that didn't stick. The audit is built for exactly that gap.
For audits, yes — every report comes from me directly. For implementation work after the audit, builds are delivered through a vetted technical partner under my supervision. You have one accountable point of contact throughout, which is me. The audit always sits with me; the build, when there is one, is supervised by me.
Then I'd rather wait until you can bring the managing partner into the room. The audit is built for the principal who owns the firm and has the authority to act on three concrete 30-day recommendations. Without that person at the table, the report doesn't move.
Where it's useful, yes — I'll name tools (and combinations of tools) that fit the firm's stack, with rough costs. Where I think the right answer is to do nothing or use what you already have, I'll say that instead. The "what I'm not recommending" page is often where the firm's existing software gets defended.
Ten in the current cohort. When those are complete I'll publish anonymised observations across the cohort and open the next round. If the cohort is full when you reach out, I'll either add you to the next one or tell you and step away — I'd rather run fewer audits well than more poorly.
Audit-eligible if you come inbound, but I'm not actively marketing to those verticals yet. The current cohort is independent UK accountancy firms. The recruitment and agency versions will follow in 2027 once the accountancy case studies are in place.
Pick a time. Four questions, a mutual yes-or-no at the end. If we're a fit we'll book the audit on the same call. If we're not, I'll tell you straight and point you to whatever I think would actually help. No follow-up, no list.