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.
38 articles
The KPIs that built successful accounting firms — utilization, AHR, profit per partner — were designed for a world where humans do all the work. When AI becomes a production participant, those metrics aren't wrong. They're just no longer enough. Here are three metrics I think matter now — and I want you to challenge me on every one of them.
AI compressed the cost of solo founding a CAS practice to near zero. Your best people know this. The question isn't how to prevent departures — it's whether your firm offers something a solo practitioner can't replicate. Four things make the difference, and all of them are buildable.
Your review queue doesn't distinguish between files that need your judgment and files that are clean. So you review everything — and become the bottleneck. The fix isn't better delegation. It's building systems that tell you where your attention actually matters, so you can stop being the dragnet and start being the editor.
A Harvard study found individuals with AI were 3x more likely to produce top-10% ideas — not because of tool skills, but because AI gave them the confidence to act on their instincts. Your team has the same instincts. The question is whether your firm builds the conditions where confidence follows — or watches your best people build it somewhere else.
Anthropic surveyed 80,508 AI users. The number one fear isn't job loss — it's unreliability. AI fails the same way your staff does. The difference: corrections become permanent rules.
Intuit and Anthropic announced a multi-year partnership putting QuickBooks inside Claude. The MCP server is already public on GitHub — a developer expanded it to 143 tools. But the depth question that's limited every platform integration still applies, and the Enterprise Suite demo isn't what your QBO clients will get.
Intuit just connected QuickBooks directly to Claude — your clients can now build AI agents with their own financial data, no accountant required. Plus every AI company is converging into every other AI company, Jeff Bezos is raising $100 billion to rebuild manufacturing with AI, and 81,000 Claude users just confirmed what we published on Wednesday.
98% of accounting firms use AI. Only 21% have a policy governing it. Your staff is already putting client data into tools you haven't vetted — and the line between safe and dangerous isn't where most practice owners think it is. Here's what's actually risky, what's safe right now, and the one thing you should implement this week.
Before every coaching call, I check with my bookkeeping team to get current on the client's status. That's not preparation — it's a tax. Every CAS practice pays it dozens of times a day: the invisible cost of moving context between people so someone can do the actual work.
Sequoia Capital says tax advisory is 80-90% intelligence work — rule-following, not judgment. Nate Jones argues most of what we call judgment is rules we've internalized so deeply we forgot they were rules. If they're right, the moat accountants keep defending on LinkedIn is real — but much narrower than the profession assumes.
AI matches experienced professionals 70-83% of the time. The profession's biggest complaint — that AI gets things wrong — is actually a description of where your professional value lives. The 17-30% where you catch the error, push back, and explain why? That's judgment. And it's the one skill nobody in the profession is developing deliberately.
The research was valuable. What happened next was more valuable — a 10-minute reflective conversation with the AI that turned a one-off project into a reusable methodology anyone in the firm can follow. That's context engineering. And it's how you stop losing what you learn.