Founder of Fuel Accountants and The AI Accountant. With over 15 years in practice management and technology consulting for accounting firms, Peter is passionate about helping accountants embrace AI to transform their businesses.
70 articles
A quiet news week with a loud through-line. Intuit cut its coordination layer, KPMG put Claude in front of 276,000 people, and the ICAEW finally put numbers on the profession's AI blind spot — and all three say the same thing: the tool was never the hard part. The system around it is, and building it is leadership work.
PE firms just bought the AI deployment layer — Goldman, Blackstone, and Hellman & Friedman backed Anthropic's enterprise JV the same week OpenAI Deployment Co finalized and Tomoro got acquired. Your SMB clients are about to want the same thing. The accountant has the natural position to deliver it — but only if they've operated it themselves first.
Anthropic just put a month-end close agent inside your client's QuickBooks. The close packet it produces is designed to land on your desk. The question is whether your engagement is priced for what the agent replaced — or for what it can't do.
Anthropic shipped Claude for Small Business on Tuesday with Intuit's on-record endorsement and 15 ready-to-run workflows that map onto every line of a small CAS practice's bottom-tier invoice. Two days later, Synthetic raised $10M for the same bet at $49 a month. Add Xero's FY26 results, XeroForce, and PwC's new Anthropic-anchored Office of the CFO, and the diagnostic question this week leaves unanswered is the one your engagement letter has to answer: who pays for quality control on the agents' work?
A working agent is a one-off; a practice runs on systems. Three things have to happen for a single agent to become practice infrastructure — it persists across time, ports across clients, and composes with other agents into chains. This piece names the operational disciplines, the platform differences, and the manual chain that lets most CAS practices start today.
Three accounting professors at UT Austin spent eight months watching 2,500 KPMG employees use AI. Only 5% reached sophisticated use — and the four behaviors that separate them aren't a new skill. They're professional skepticism with new vocabulary, and small CAS practices have structural advantages over the Big Four to deploy them.
Anthropic and OpenAI announced $11.5B in same-day enterprise joint ventures and admitted on stage that selling models isn't enough — the bottleneck is organizational absorption. The CAS-shaped agents Anthropic shipped the next day aren't pointed at your QBO clients, but the platform vendor that does reach them launched its second major release in eight days. And three large public companies redesigned their org charts around AI-native operating models in the same calendar week.
You have two pieces of a working agent — an instruction from Part 1, a context file from Part 3. The third piece is the example, and the trick is you don't write it. Run the agent without one. Fix the instruction or the context — not the output — until what comes back is what you'd send. That output is your example.
On April 17, Anthropic launched Claude Design — free on every paid Claude plan, turning prompts and source data into branded slides, PDFs, and reports. For CAS practices, three patterns are worth your time this quarter — and one principle decides whether the tool actually pays off.
Sage and PwC's Beyond the Black Box research put a number on AI's hidden cost: finance professionals spend 12.9 hours a week reconstructing and explaining AI outputs, with 26% of AI time savings lost to verification. The 71% rejection rate the headlines led with is the future cost; the 12.9 hours is what your firm pays now — usually out of partner time. The fix is to stop shipping AI deliverables alone and start shipping them with a research memo and a cross-model verification report.
Intuit attached a public price tag to AI-driven close work at $8 per client per month, Sage rebuilt its portfolio around agents at Sage Future, and the standalone tools you pay for every month shipped AI rebuilds the same week. The bookkeeping service line is being absorbed into the platform subscription in front of you. The 71% trust gap from Sage's IDC research is the only thing slowing it down — and trust is a lagging constraint, not a permanent one.
For months I've been writing about client context like its meaning is obvious. It isn't. This is the article where we stop talking about context engineering in the abstract and actually build one — a single document that makes every agent you run for a client better, and that lets you take the same agent to every other client just by swapping files.