This isn't a fringe position. Wipfli, RSM, Karbon, Schellman, Fieldguide — firms across the market cap spectrum — are all modeling revenue growth against flat or declining headcount. Accounting Today documented the pattern in April 2026: tech spending is now outpacing people spending at firms that were once defined by their billable-hour pyramid.

If you own a 10 to 30-person accounting firm, here's what you need to understand about what this shift actually looks like — and what the early movers built.

The Wipfli Three-Week Build

Wipfli is a Top 20 accounting firm. They're not small. But the thing they built that matters to you doesn't require a Top 20 budget.

In three weeks, Wipfli deployed a working agentic AI system for legislative monitoring. The system uses two AI agents in sequence: a researcher agent that monitors and gathers legislative updates, and a curator agent that synthesizes and packages the output into a deliverable for mid-market clients. No proprietary infrastructure. Off-the-shelf tools. Three weeks from concept to deployment.

The practical translation for a 10-person firm: that same architecture — researcher agent plus curator agent — can monitor IRS guidance, state tax law changes, or regulatory updates relevant to your client base. The human reviews the output. The agents do the watching, gathering, and drafting.

The question to ask is not "can we afford to build this?" The question is "what monitoring or research work is currently taking us 5-10 hours per month that an agentic workflow could handle?"

What Schellman Built

Schellman is a cybersecurity and compliance assessment firm — attestation services, SOC reports, ISO certifications. They built an agentic AI system for compliance training: automated delivery, tracking, and quality control for staff education on SOC and ISO standards.

The parallel for a small accounting or advisory firm is direct. Think about where your team spends time on structured, repeatable information delivery: client onboarding, engagement letter customization, staff training on new standards, quarterly client education reports. Any of those is a candidate for an agentic workflow.

Schellman's system handles the repetitive content layer. Humans review outputs and handle judgment calls. That's the model.

What "Flat Headcount, Growing Revenue" Actually Means

When a firm commits to this structure — tech-first, headcount-flat — the math changes in three ways:

1. Revenue per team member goes up. If your team handles 20% more client work without adding staff, every revenue dollar costs less to produce. Margins improve. You can take on more clients, raise prices, or both.

2. The hiring bar rises. When you're not hiring for volume — when every new hire needs to be a net addition to judgment and client relationships, not a body for processing — you hire differently. Fewer, better people. That's a talent strategy, not just a cost strategy.

3. The breakeven on AI investment is short. A Wipfli-style agentic build using Claude or GPT-4o plus a workflow automation tool like Make or Zapier costs roughly $200-500 per month in tooling, plus 20-40 hours of setup time. If it replaces 10-15 hours per month of staff work, payback is under 90 days. Most implementations pay back faster.

What This Doesn't Mean

This shift does not mean AI replaces accountants. The pattern at every firm making this transition is the same: AI handles structured, repeatable, research-heavy tasks. CPAs handle client relationships, judgment, review, and advice.

What changes is the ratio. One CPA, with the right workflows, can serve more clients and produce more deliverables than they could two years ago. The firms that figure this out first won't necessarily grow their headcount — but they will grow their revenue per person.

Where to Start

You don't need a three-week sprint to build your first agent. You need one workflow.

Pick the task in your firm that is:

• Repeatable (you do it at least weekly)
• Research-heavy or draft-heavy (it requires gathering information and formatting it)
• Currently eating hours that your best people resent spending

That's your first candidate. For most accounting firms, the highest-value starting points are:

• IRS notice triage and response — classify notices, research code sections, draft response letters
• Legislative monitoring — track state tax law changes relevant to your client base
• Client-facing summaries — convert complex analysis into plain-English client communications

Build the first one. Measure the time savings. Then replicate the model.

The firms making the tech-over-headcount shift didn't start with a strategic transformation initiative. They started with one workflow that worked. Then they built another.

---

The Crossing Report covers AI adoption for professional services firm owners every week. Subscribe here.