About Peter
About Peter McCarroll
I didn't want to be the accountant watching from the sideline while a tank ran over my business.
I run a CAS practice. I've been in public accounting for over two decades. And when AI started reshaping the profession, I decided to figure it out from the inside — not wait for someone to explain it to me.
The Moment Everything Changed
May 2024. After 20+ years in accounting, I got worried.
I still remember it clearly. I'd been running my practice at Fuel Accountants for over 10 years, and I'd worked in public accounting for more than a decade before that. We offered the full suite — bookkeeping, accounting, tax, Profit First implementation, and advisory. A growing practice. Good clients. A team I was proud of.
Then the noise about AI got impossible to ignore. ChatGPT. Machine learning. Agentic workflows. I wasn't just hearing about it — I was starting to see what it could actually do. And it scared me.
Not because I didn't believe in it. Because I did. I could see where this was heading, and I knew that the compliance work we'd built our revenue on — the bookkeeping, the bank recs, the tax prep — was about to get compressed. Hard.
I didn't want to be the accountant sitting on the side of the road crying because someone had just run over my business in their tank. AI is the tank. I decided then and there that I needed to learn everything I could — not to protect what I had, but to build what comes next.
The Problem
The training I needed didn't exist.
Everything available fell into one of two categories — neither of which worked for accountants.
Too Generic
Built for marketers and salespeople
Most AI training was designed for people writing blog posts and email sequences. The examples didn't translate. The workflows didn't apply. And the instructors had never reconciled a bank statement in their lives.
Too Technical
Built for developers and data scientists
The serious AI education assumed you could write Python and spin up APIs. Useful if you're an engineer. Useless if you're trying to figure out how to apply AI to a monthly close cycle or a set of workpapers.
Too Shallow
Built for accountants, but barely scratching the surface
The training that did target accountants stopped at "here's how to write a prompt." No workflow design. No context engineering. No path from experimentation to operational change. It was awareness-level content dressed up as implementation.
Nothing spoke to the way accountants actually work. Nobody was building AI training around chart of accounts management, client deliverable workflows, year-end processes, or advisory prep. Nobody was addressing the specific constraints of CAS practices — the platform lock-in, the compliance requirements, the client expectations.
So I built it myself.
What I Built
An AI system designed from inside a CAS practice.
I started the way most of us do — experimenting. Prompt by prompt, workflow by workflow, figuring out what AI could actually handle inside a real practice with real clients and real deadlines. Some things worked immediately. Others took months of iteration. A few ideas I threw out entirely.
What emerged wasn't a set of tips or a prompt library. It was an operating system — a structured approach to identifying where AI fits in CAS workflows, training people to use it consistently, and measuring whether it's actually changing anything.
Along the way, I connected with Steve Cunningham at Humans+Agents, whose AI methodology had been deployed at Meta, Microsoft, PwC, and Deloitte. What I'd been building inside my own firm aligned with the same principles he'd proven at enterprise scale. We combined forces — his methodology, my CAS-specific implementation experience — and The AI Accountant was born.
Tested in a real practice
Every framework, every workflow, every training module was built and tested inside Fuel Accountants before it was offered to anyone else. This isn't theory — it's what we run.
Enterprise methodology, CAS application
The Humans+Agents operating system has trained professionals at four of the largest companies and professional services firms in the world. We've adapted it specifically for mid-market CAS practices with 5 to 50 staff.
Practitioner-led, not consultant-led
I'm not a tech consultant who read about accounting. I'm an accountant who figured out AI. That distinction matters when you're training a team that needs to see someone who understands their actual work.
By the Numbers
Built on experience, not credentials alone.
Why This Matters
The profession is changing. The question is whether you're leading the change or reacting to it.
Compliance work is compressing toward zero margin. That's not a prediction — it's already happening. Clients are pushing back on fees that were standard two years ago. New entrants are offering bookkeeping at prices that don't make sense unless they're using AI to deliver it.
But here's what I've learned: the firms that figure this out don't just survive — they grow. They shift from volume-based billing to value-based pricing. Advisory becomes the product, not the upsell. Their teams work faster, deliver more, and spend their time on work that actually requires human judgment.
That's what I'm building toward — with my own firm and with every practice that works with The AI Accountant. Not AI for its own sake. AI as the engine that lets a firm that serves clients well stay competitive, grow capacity without adding headcount, and charge for the thing that actually matters: the advice, the accountability, and the relationship.
Let's Talk
Want to know what AI could look like inside your practice?
I'll give you an honest assessment — where the opportunities are, what's realistic, and whether our programs are the right fit. No pitch. Just a conversation between practitioners.