Accounting firms are adopting artificial intelligence faster than at any point in recent memory, but a new Rightworks survey suggests the financial payoff depends less on the tools themselves than on what firms do around them. AI adoption across firms has surged 78%, and active users report significantly higher revenue per employee. Yet most firms still operate without formal AI usage policies. That gap between capability and control may prove the defining operational risk of the next two years.
The Adoption Curve Is Steep. The Policy Layer Is Thin.
The headline finding is straightforward: usage is climbing quickly. The less comfortable finding is that governance is trailing behind. Many firms have moved from experimentation to daily use of drafting assistants, research tools, and workpaper automation, but few have written rules describing what staff may enter into those systems, how outputs are reviewed, or who owns the resulting judgment.
AI adoption across accounting firms has surged 78%, yielding significantly higher revenue per employee, even as most firms still lack formal AI usage policies. (Rightworks survey, as reported in the primary survey coverage)
For practitioners, this is less a technology story than a supervision story. A tool that drafts a memo or summarizes a trial balance can accelerate work, but it also introduces questions about confidentiality, reliance, documentation, and accountability that existing engagement methodology may not fully answer.
Reading the Revenue Signal Carefully
The most eye-catching figure is that active AI users report 39% higher revenue per employee. That is a meaningful signal, but it should be interpreted with discipline. Firms that adopt AI early may also be larger, more digitally mature, or more willing to invest in technology and training, all of which could independently lift productivity. Revenue per employee also reflects pricing decisions, staffing mix, and client selection, not just efficiency.
The useful question for managing partners is not whether AI raises revenue in the abstract. It is where in the service delivery chain the productivity gain actually lands, and whether the firm captures it.
| Signal | What It Suggests | What Firms Should Verify |
|---|---|---|
| 78% adoption growth | AI has moved from pilot to standard workflow in many practices | Which tools are in use, and whether they are sanctioned or shadow-deployed |
| 39% higher revenue per employee (active users) | Productivity gains are showing up in top-line output | Whether gains come from speed, scope, pricing, or mix |
| Most firms lack formal policies | Governance is lagging capability | Written acceptable-use standards, data handling rules, and review requirements |
Where the Margin Actually Goes
If AI reduces the hours required for a compliance deliverable or advisory analysis, a firm billing by the hour may see revenue fall even as productivity rises. This is the classic efficiency dilemma, and it is why pricing architecture is now part of the AI conversation. Infinite Ties recently launched a growth and profitability engine designed for client advisory services practices that helps firms evaluate delivery costs, establish optimal fixed-fee pricing, and protect engagement margins.
That kind of tooling reflects a broader shift. Firms that treat AI as a cost-cutting measure may compress fees. Firms that price around outcomes and defined deliverables are better positioned to keep the efficiency gains they create.
The Policy Gap Is a Liability, Not a Formality
Regulators are increasingly focused on how firms manage process and documentation. The PCAOB recently proposed targeted amendments to AS 6101 covering comfort-letter conventions and announced a restructuring of its Division of Enforcement into five specialized units. Whatever the specific AI implications of those moves, the direction is clear: regulatory attention is becoming more specialized and more granular.
In that environment, an undocumented AI workflow is a potential audit trail gap. Firms should be able to answer basic questions: which tool produced a given analysis, what data was provided to it, who reviewed the output, and how that review was evidenced in the file.
Minimum Elements of an AI Usage Policy
- Approved tools list: Enterprise-sanctioned platforms only, with clear rules for personal accounts
- Data classification: What client information may and may not be entered into AI systems
- Review standards: Required human review and sign-off before AI-assisted work product is relied upon
- Documentation: How AI involvement is recorded in engagement files
- Training and accountability: Who owns the policy and how staff are trained to apply it
Market Signals: Consolidation and AI Distribution
Two other developments point to where the market is heading. Larson Gross completed a merger with Petersen CPAs and Advisors in Washington State, its fifth acquisition of the year, illustrating how scale-driven consolidation continues among top firms. Scale can fund technology investment, but it also raises the stakes of inconsistent governance across offices and service lines.
On the enterprise side, BlackLine made its accounting tools and Verity AI capabilities available through Anthropic's Claude Marketplace, aimed at enterprise finance teams. When AI tools are distributed through marketplaces and embedded in client finance systems, CPA firms will increasingly encounter AI-generated schedules, reconciliations, and close documentation from their clients. Reviewing that output will become part of routine assurance and advisory work.
Tax and Disclosure Work Is Also Changing
Technology pressure is not limited to firm operations. At the AICPA National Tax Conference, National Taxpayer Advocate Erin Collins highlighted upcoming Tax Pro Account updates that will let practitioners submit signed powers of attorney electronically. Faster digital authorization will reduce administrative friction, but it also means more client authorizations will flow through digital channels that firms must govern with the same rigor as other data systems.
Similarly, the SEC's proposal to modernize executive compensation disclosure rules under Regulation S-K Item 402, with a focus on materiality and streamlined tabular reporting, will change the drafting and review workload for firms serving public company clients. Lighter tables may reduce some manual effort, but judgment about materiality becomes more central.
A Practical Roadmap for the Next 90 Days
- Inventory actual AI use. Survey staff on which tools they use, including unsanctioned ones, before writing policy.
- Adopt a written AI usage policy. Cover approved tools, data handling, review requirements, and documentation standards.
- Map gains to pricing. Identify which services have become faster and decide whether to move toward fixed-fee or outcome-based models.
- Build review into workflow. Make human sign-off a defined step, not an informal habit.
- Train and assign ownership. Name a partner or technology lead accountable for AI governance across service lines.
Looking Ahead
The Rightworks findings suggest that AI is already producing measurable economic results for accounting firms. The competitive question is shifting from whether to adopt to how well a firm can convert adoption into durable, well-governed, appropriately priced work. Firms that build policy, pricing discipline, and review standards now will be better placed to capture productivity gains without absorbing the regulatory and reputational risks that follow unmanaged technology. Those that wait may find that their competitors have already converted efficiency into a sustainable advantage.
