The results so far: approximately 20% immediate time savings from the platform migration alone. His long-term projection: 30–50% efficiency once the full client base migrates.

He summarized it the way a real firm owner summarizes it: "The less time we spend on repetitive tasks, the more time we can spend helping clients make better business decisions."

That's the whole story of AI adoption at a small professional services firm. Not a transformation. Not a moonshot. Less time on the work that doesn't require your brain, more time on the work that does. And a number — 20% — that's real enough to be credible and meaningful enough to matter.

(Source: CPA Practice Advisor, April 15, 2026)

---

Why This Case Study Matters

The Tonkin story is not remarkable because the numbers are large. It's remarkable because the firm owner is not a technology early adopter, the numbers are conservative, and the efficiency gain is already real — not projected.

Most AI case studies that circulate in the accounting profession come from large firms with dedicated technology staff, or from AI vendors with incentive to publish the highest possible numbers. They feature 10-person pilot teams, phased rollouts, and headline percentages that are difficult to translate to a 3-person family firm.

The Tonkin case is different: a family-run bookkeeping practice, a straightforward platform switch, and a 20% time savings figure that's the floor, not the ceiling.

If you've been watching AI case studies from the sidelines and thinking "that's not my firm" — this one is closer to your firm.

---

What Digits Actually Does

Digits is an AI-native accounting platform built around a simple premise: the work of bookkeeping — categorizing transactions, reconciling accounts, generating reports, flagging anomalies — is almost entirely pattern recognition. AI is better at pattern recognition than humans are. Humans are better at judgment, interpretation, and client relationships than AI is.

Digits automates the pattern recognition work. The accountant's job becomes reviewing what the AI has categorized, catching the exceptions that require judgment, and spending the remaining time on the interpretation and advisory work that clients actually value.

For a firm like Steve Tonkin & Company — recurring bookkeeping and controllership engagements for small business clients — this changes the staffing math. Less time per client per month on the transactional work. More capacity for advisory conversations. Or the same advisory quality with more clients.

Digits also announced outcome-based pricing in April 2026: firms pay only for "zero-touch transactions" — transactions where AI handled the classification without human intervention. The billing is based on what AI measurably replaced. Not on subscription seats, not on upfront commitments. You see exactly what you're paying for.

---

The Migration Is the Work

The 20% immediate savings figure represents a partial client migration. Not all of Tonkin's clients are on Digits yet.

This is the honest version of every AI platform story: the efficiency gains scale with how much of your workflow you've actually moved. The platform migration is the work.

For a small accounting firm, the migration looks like this:

Month 1–3: Connect your first batch of clients to Digits. These are typically your most straightforward bookkeeping clients — small businesses with standard transaction types and manageable volume. The AI needs 60–90 days of transaction history to establish reliable categorization patterns for each client.

Month 3–6: You'll start seeing the 15–25% time savings on these early-migrated clients as the AI pattern recognition matures. Your staff spends less time on data entry and categorization for these accounts; they spend more time on review and advisory.

Month 6–18: Migrate remaining clients. More complex clients take longer — higher transaction volume, unusual industry patterns, legacy system integration. Some clients may resist migrating because they're comfortable with your current process. The efficiency gains from partial migration plateau until you move the full book.

Post-migration: The 30–50% efficiency range Tonkin projects reflects a fully migrated client base where AI is handling the bulk of transactional classification on every engagement. At this point, the capacity freed is substantial enough to either serve more clients or deliver significantly more advisory depth to the same client base.

The migration is not a one-day decision. It's a 12–18 month project for most firms. But the gains start in month three, not month eighteen.

---

What You're Actually Trading

The objection to AI-native accounting platforms is usually one of three things:

"I don't trust AI to categorize my clients' transactions accurately." Digits provides an audit trail of every categorization decision — AI or human. You review AI categorizations before they close. The trust is built through the verification process, not assumed on day one. By month three, you'll know exactly where the AI is reliable and where it needs your review.

"My clients won't want their work done by AI." Most clients don't care how the books are kept. They care that the books are accurate, the reports are timely, and their accountant shows up when they have a question. If AI enables you to respond faster and advise more proactively, that's a better client experience, not a worse one.

"The migration is too much work for a firm my size." This is the most legitimate concern. Migration takes time. The right answer is not to migrate everything at once. Start with three clients — your most straightforward recurring bookkeeping accounts. See how the AI performs, understand the exception-handling workflow, and build confidence before you migrate the full book.

The Tonkin story is a 20% savings figure after a partial migration. That's the floor of what's available. The ceiling is a fundamentally different kind of firm — one where the owner spends client time on interpretation and planning, not reconciliation and data entry.

---

The Question to Ask This Week

If you're a bookkeeping or accounting firm owner who has been watching AI adoption from the sidelines, the Tonkin case study reduces the question to one thing:

What is 20% of my time currently spent on worth to me?

If the answer is "I'd rather spend that time on client advisory conversations," that's the business case. If the answer is "I'd rather take on three more clients without adding staff," that's the business case. If the answer is "I'd rather work 20% fewer hours and keep my revenue the same," that's also the business case.

The efficiency exists. What you do with it is your decision. But it's available now, not in two years.

(Sources: CPA Practice Advisor, April 15, 2026; Digits.com)