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.
117 articles
Jason Staats listed the six reasons firms get stuck on AI — every one written for the partner who signs the cheque. But if you're the one who got told to "figure out AI," the blockers look completely different. Here's what you can do without anyone's permission, and the short list to ask your firm for.
On Tuesday, Anthropic released the most capable AI model ever made public. By Friday, the US government had switched it off worldwide. Add OpenAI retiring models on its own schedule, Google force-defaulting Gemini, and QuickBooks killing classic reports — and the week's lesson is hard to miss: you don't own your AI supply chain. Here's how to build for it.
AICPA leadership opened Engage 2026 by naming AI the top issue for CPA firms of every size. After four days on the floor — AI-native ledgers, closed walls, and a profession aiming at 2030 — here's why the real deadline is 2028.
In the last few months, agents running on Opus 4.8 and GPT-5.5 changed what AI can do — but the training most firms will buy to keep up is still a 2024 prompt-engineering course. The skills that actually build an AI-native team are learned through reps in your own work, and the most valuable may not even count as CPD. Here are the six questions to grade any AI training before you book it.
Ramp shipped an "AI operating system" built for accounting firms, Karbon and Firm360 opened MCP sockets to whatever AI you bring, Thrive put $1 billion on the table to buy firms like yours — and 69% of tax pros told a surveyor the hourly billing model is ending. Not one major model release all week. The fight has moved up the stack, to who holds your firm's encoded knowledge and who controls the door.
YC's AI stack works because they gave the agent direct access to their data. Most CAS firms have done the opposite — locked AI out of everything and then wondered why it isn't very smart. The real question for your firm isn't whether to give AI access. It's how much, of what, with what controls.
OpenAI and Thrive just shipped a self-improving tax agent at 97% draft accuracy — built by six months of OpenAI's forward-deployed engineers and researchers embedded across a 30-firm network. It isn't a service OpenAI sells; it's what happens inside a company OpenAI owns equity in. But the loop the agent runs on sits one layer down the stack, fits in your sprint, and is the only version that doesn't turn your firm into someone else's training data.
You've inherited that workbook — beautiful structure, validated, and wrong every month nobody caught it. AI now ships polish in seconds, and in CAS, polish is what your clients pay for as a proxy for trust. Here's the one prompting habit, and the design principle underneath it, that separates Champions from Copilot users.
OpenAI shipped a working tax agent at 97% draft accuracy — but it took a custom-built product to get there, not a ChatGPT login. Kick opened its ledger to whichever AI you bring, and made it work across all your clients at once. Microsoft canceled Claude Code over runaway costs, a CFO ate a half-billion-dollar surprise bill, and Opus 4.8 got noticeably more honest. AI implementation is hard, can get expensive, and produces powerful results when done well — the firms thinking 12 to 18 months ahead are the ones who'll come out on top.
Six layers — firm identity, tech stack map, client registry, workflow templates, agent slots, monitoring. Four of them have real value with zero AI agents running. Most CAS firms don't have any of them in deliberate form, which is why their AI investments keep producing chatbots wearing the firm's logo instead of real changes to how the practice operates.
Ian Crosby, the founder behind Bench, just launched an AI service that does a small business's books for $49 a month. The reflex is to argue AI isn't good enough yet — but that's the wrong question. As production automates and the data stops arriving pre-trusted, the durable human role moves to quality control: catching what the AI gets wrong, standing behind the numbers, and turning that into advice.
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.