Running an Accounting Practice Solo with AI: A CPA with Zero Employees and an Accountant Who Doubled Her Firm
A Virginia CPA runs his firm with no employees and has AI draft tax workpapers in minutes. An Adelaide accountant doubled her revenue and client base without hiring staff. Here is what each one hands to AI, and what their accounting experience still has to do.
Founders in this article
- Sam LeonUnited States
- Emma FabbroAustralia
For a long time, an accounting firm of one meant a ceiling. One person could only prepare so many returns and close so many books, so growing meant hiring staff and managing them.
Two practitioners show how AI changes that math. One is a CPA in Virginia who runs his firm with no employees and has AI do the first draft of his tax workpapers. The other is an accountant in Adelaide who doubled her revenue and client base over a few years without adding staff.
This article is for accountants and bookkeepers who are thinking about going out on their own. We look at what each of them hands to AI, and then, in a separate section, at what their professional experience still has to do.
Sam Leon: a CPA firm with zero employees
Sam Leon runs The Millennial CPA, a digital tax practice based in Richmond, Virginia. Before starting it, he worked at larger accounting firms. Today his firm has one office and no employees, and by the account of accounting podcast host Blake Oliver, about 70 percent of its budget goes to technology. The firm appears on Accounting Today's 2026 list of Best Firms for Technology.
How he uses AI
- AI prepares, he reviews. He uses projects in Anthropic's Claude to turn client documents into tax workpapers. Workpapers that used to take a person three to five hours to build from financial statements now take AI about five minutes, according to Oliver.
- Faster year-over-year comparisons. Comparisons with the prior year, which once took an afternoon, are done in a fraction of the time.
- Review is the new bottleneck. On The Accounting Podcast he discussed where AI still falls short and why reviewing its output, not producing it, is now the slowest step.
- Hiring only when AI hits its limit. His approach to staffing, as Oliver describes it, is not to hire until AI can no longer handle the work.
What it became
He still applies his own professional judgment to every return, but spends far less time on document collection, data entry and workpaper preparation. Alongside the firm, he is building software for tax and accounting firms, according to his author profile at Accounting Today.
Emma Fabbro: an accountant who doubled her firm without hiring staff
Emma Fabbro runs Fusion Accountants in Adelaide, South Australia. Over the past three to four years, her firm roughly doubled in both revenue and client capacity. She did it without building a staff team, working instead with a small group of specialist contractors and relying heavily on her accounting software.
How she uses AI
- Lead replies drafted for her. Her Gmail is linked to Customer AI, a feature of Intuit Accountant Suite, which spots new client leads and drafts responses. In her words: "The AI suggests the draft, I provide the human approval."
- Routine practice work automated. Client reminders, bank feed categorisation and meeting scheduling run inside the same software.
- Client data flows in on its own. Clients capture receipts and mileage in the QuickBooks mobile app, and that data arrives in her system without her chasing paperwork.
- Changing tools quickly. She argues that a firm her size can test a new tool on a Friday and have it in use by Monday, which large firms cannot.
What it became
The time saved lets her step away from the office without the practice stalling. She brings in contractors with specialist expertise for work that needs human judgment, and spends her own time on advice for small business owners.
What accounting experience gave them that AI could not
This section is our analysis, not a claim made by either practitioner.
AI handles the preparation in both cases. What makes it safe to rely on comes from their training and experience.
- Knowing what a correct result looks like. Leon can let AI build a workpaper in minutes because he can tell, in review, whether it is right. Without that skill, faster drafts would only mean faster mistakes.
- Knowing where the hours go. Both targeted the work they knew ate their time: workpaper preparation for Leon, client chasing and admin for Fabbro.
- Keeping advice human. Both draw the same line: AI drafts and flags, and a professional decides what to tell the client.
The two cases also show two shapes of a solo accounting practice:
| Sam Leon | Emma Fabbro | |
|---|---|---|
| Firm | The Millennial CPA, Richmond, Virginia | Fusion Accountants, Adelaide |
| Team | No employees | No staff team; specialist contractors |
| Main AI tool | Claude (projects) | Customer AI in Intuit Accountant Suite |
| What AI handles | Turning client documents into tax workpapers, year-over-year comparisons | Drafting replies to leads, client reminders, bank feed categorisation, scheduling |
| What stays human | Review and professional judgment on every return | Approving drafts, advice to clients, specialist work through contractors |
| Reported result | Tax prep about five times faster | Revenue and client capacity roughly doubled in three to four years |
What you can copy
These are our suggestions, drawn from the two stories.
- Start with preparation, not advice. Let AI draft workpapers, categorise transactions or draft emails. Keep the judgment calls yourself.
- Plan for review time. If AI makes preparation five times faster, your review becomes the limit. Build a checklist for checking AI output.
- Use the AI inside the software you already pay for. Fabbro's gains came largely from features in her existing accounting suite, not from a separate tool.
- Grow with contractors before employees. Both kept the firm small: one with no staff at all, the other with specialist contractors for the work that needs a person.
- Mind client data. Decide which client documents may go into which AI tool before you start, and use paid plans that suit professional work.
Notes on the sources
The facts about Sam Leon come from a blog post and podcast episode by Blake Oliver, host of The Accounting Podcast, and from Accounting Today. The figures on time saved are as reported there. The article about Emma Fabbro was published in Accounting Times under the byline of Intuit QuickBooks, the maker of the software she uses, so it is sponsored content. Figures are as of the dates of those sources in 2026.
Sources
- The Solo Firm Isn't New. Its Ability to Scale Is., Blake Oliver (2026)
- How Sam Leon Uses Claude to Prep Tax Returns 5x Faster, The Accounting Podcast episode page (2026)
- The 2026 Best Accounting Firms for Technology, Accounting Today
- Sam Leon author profile, Accounting Today
- Beachside bookkeeping: The Adelaide sole practitioner proving AI can double your revenue, Accounting Times, by Intuit QuickBooks (sponsored), March 24, 2026
Related articles
A Former Recruiter Who Now Teaches Recruiters to Use AI: Vanessa Raath and The Talent Hunter
Vanessa Raath spent years as a recruiter and head of talent acquisition before starting her own sourcing training business in South Africa in 2019. After ChatGPT launched, she began teaching recruiters how to use AI on their desks, and she has trained more than 8,000 recruitment professionals in over 100 countries. Here is the workflow she uses to turn a CV, interview notes and a LinkedIn profile into a client-ready summary in under five minutes.
A Spanish History Teacher Built an AI Platform for Talking with Historical Figures: Julio Candela and Habla con la Historia
Julio Candela teaches geography and history at a secondary school in Elche, Spain. In October 2023 he launched Habla con la Historia, a website where students talk with AI versions of historical figures and watch them debate. It drew more than 83,000 users from 25 countries in its first half of 2024.
A Freelance Ads Consultant Measured What AI Saves Him: Lionel Fenestraz and 14 Hours a Week
Lionel Fenestraz is a freelance Google Ads and Meta Ads consultant in Alicante, Spain. Over six months he measured, task by task, how much time Claude saves him: monthly reports, client briefs, search term analysis and SEO audits. By his count it adds up to about 14 hours a week, which he used to take on three new clients without hiring. Here is what he hands to AI, and what stays with him.
A Solo Counselor Who Has No One to Audit His Notes: How Tom Schlosser Uses AI for Clinical Documentation
Tom Schlosser is a licensed professional counselor who runs a solo private practice south of Pittsburgh. Writing progress notes used to take him hours of catch-up and repeated checking. AI built into his electronic health record now drafts and polishes his notes, saving him about an hour a week. Here is what he hands to AI, and what stays with the clinician.