Audit Fees Are the Next Pricing War — What Small CPA Firms Should Do Before It Arrives
Published March 15, 2026 · By The Crossing Report
Audit Fees Are the Next Pricing War — What Small CPA Firms Should Do Before It Arrives
KPMG just proved the thesis. OnlyCFO published the analysis. Bloomberg Tax confirmed it's already happening at multiple large firms. And the pattern is clear enough that small CPA firm owners need to act on it now — before the first client asks the question.
The thesis: AI is compressing the hours it takes to complete an audit engagement. When one competitor's cost per engagement drops while yours stays flat, they can price below you on the same scope and still make margin. The fee squeeze doesn't start with your clients asking — it starts with a competitor submitting a proposal you can't match.
Here's what's happening and what to do about it.
What KPMG Proved
KPMG has been the most public of the Big Four about applying AI to audit workflows. Their AI integration targets the non-judgment work that makes up the largest share of audit labor hours: transaction classification, document review and categorization, variance flagging, reconciliation checks, and control testing on high-volume, low-complexity items.
OnlyCFO's 2026 analysis documents the mechanism precisely: AI doesn't replace the auditor's judgment on complex items, material estimates, or going-concern assessments. What it does is collapse the time spent on the deterministic work that happens before and around those judgment calls. A transaction sample that took a staff accountant four hours now takes an AI tool 20 minutes, with the auditor reviewing the flagged exceptions rather than coding the full population.
Bloomberg Tax confirmed in early 2026 that AI efficiency gains are already pushing large accounting firms to "reimagine pricing" — meaning they're fielding client questions about why audit fees haven't reflected the efficiency gains. In some cases, firms are absorbing the savings as margin improvement. In others, they're beginning to pass them through to retain clients who've done the math.
For small CPA firms: you are not competing with KPMG's audit clients. But you are competing with regional and national firms that are running the same AI playbook at a scale much closer to yours.
The Two-Phase Pressure Timeline
Phase one is competitive pressure, already underway. Regional firms that have integrated AI into their audit workflows are submitting proposals that beat your price on comparable scope. They're not discounting — they're maintaining margin while pricing at a level you can't match with your current cost structure. If you've lost an audit engagement to a regional competitor in the past 12 months and couldn't figure out why, this may be part of the explanation.
Phase two is client pressure, likely 12–24 months out. This is the wave that hits when clients who work with larger firms — or who simply read the accounting press — start asking their CPA firm directly: "We know AI is cutting audit hours at other firms. Why hasn't our engagement fee reflected that?" This question is already being asked in corporate legal and consulting. It will arrive in audit.
The window to automate before clients ask is open now. Firms that have already integrated AI into their audit workflows by the time clients start asking will have a compelling answer: "Our AI integration improved your accuracy and turnaround — here's what that delivered." Firms that haven't will have a much harder conversation.
Three Audit Workflow Categories to Automate First
Not all audit work is equally automatable. The categories with the best return for small CPA firms — high-volume, deterministic, low judgment — are also the ones where AI tools have the most traction in 2026.
1. Transaction classification and coding
Ramp's Accounting Agent (launched February 2026) reports 90%+ auto-coding accuracy on business expenses, with early clients reducing month-end close time by 3x and saving 40+ hours per month on transaction processing. For small audit practices that manage client bookkeeping before the audit, this is the first workflow to address. The AI codes, you review the exceptions.
Tools with traction at small firm scale: Ramp Accounting Agent (for clients using Ramp for expense management), BILL Accounts Payable AI, and Keeper (AI-assisted bookkeeping for tax/accounting practices). All have small-firm pricing and no enterprise minimums.
2. Document classification and intake
Audit fieldwork begins with collecting and organizing support documentation: bank statements, invoices, contracts, payroll records. AI document classification tools can sort, label, and route incoming documentation before a staff accountant touches it. This step typically takes 20–40% of early fieldwork hours on a standard engagement.
FloQast and Caseware both have AI-assisted document intake tools. Diligent's AuditAI (launched March 2026, IIA Global internal audit management) reported 70% reduction in audit administrative hours — from 120 to 35 hours per engagement in early use cases. The specific workflow: document upload → AI classification and completeness check → exception flagging for auditor review.
3. Variance analysis and anomaly detection
Variance analysis — comparing accounts, periods, and ratios to flag unusual items for auditor investigation — is the most time-intensive deterministic step in many smaller engagements. AI tools can run variance analysis across an entire general ledger in minutes, producing a prioritized list of items that warrant auditor attention. This replaces the manual formula-building and pivot table work that staff accountants typically do.
Tools: Claude Sonnet 4.6's 1 million token context window (in beta) now makes it practical to upload an entire multi-year general ledger and ask for anomaly flags across the full dataset. For purpose-built accounting AI: Thomson Reuters Checkpoint Edge and Black Ore Tax Autopilot both include variance analysis capabilities in their audit and tax workflows.
Positioning When Clients Ask
The fee pressure question will come. When it does, the correct answer is not "we lowered our fees because AI made the work cheaper." That's how you train clients to expect your rates to fall every year as AI improves.
The correct answer is: "AI improved our engagement process — faster document processing, lower error rate on the deterministic work, more time for the analysis and advisory work that requires our judgment. Our fees reflect the value of our professional judgment, not the hours of repetitive work. What you're getting is the same coverage with better accuracy and faster delivery."
Firms that frame AI as a quality improvement — same scope, better outcomes, faster turnaround — hold their rate structure. Firms that frame AI as a cost reduction invite the negotiation.
This framing requires that you can actually point to the improvement: shorter fieldwork cycles, lower error rates, faster close, more time spent in advisory conversations. That means tracking these metrics now, before clients ask, so you have the data when the question comes.
What to Do This Week
If you're an audit-focused small CPA firm (3–15 staff): Run one engagement through this question: which steps in your standard audit process are deterministic — rule-based, high-volume, low judgment? Pick the one step that consumes the most staff time. Research whether one of the tools above applies. Most have free trials or demo access without a sales call. Run the test on a current engagement with appropriate oversight. Time the before and after.
If you do tax and bookkeeping with occasional audit work: The competitive pressure is lower and later for you. But the Intuit/Anthropic partnership (rolling out to QuickBooks this spring) and tools like Black Ore Tax Autopilot and Accrual are already changing what clients expect from their accounting firm. The efficiency argument applies to your bookkeeping and tax prep workflows before it reaches your audit work.
In either case: Track your hours per engagement now. You need a baseline to measure the improvement — and to have the data ready when a client asks whether your AI adoption has changed their engagement's cost structure.
Related: The Intuit/Anthropic QuickBooks Partnership: What Accounting Firms Need to Know | PwC's Audit Automation Timeline: What Small CPA Firms Have 9 Months to Do | Diligent AuditAI: 70% Admin Reduction — What It Means for Small CPA Firms
Frequently Asked Questions
Are Big Four firms using AI to cut audit hours?
Yes. KPMG, Deloitte, and PwC have all announced AI integration into core audit workflows — transaction sampling, document classification, reconciliation review, and variance analysis. KPMG in particular has been public about AI enabling it to do the same audit scope with fewer labor hours per engagement. OnlyCFO's 2026 analysis documents the mechanism: AI reduces the hours billed to an engagement without reducing the scope covered. Clients notice this in their invoices — or start asking why they haven't.
When will AI audit fee pressure reach small CPA firms?
The timeline has two phases. Phase one — already happening — is competitive pressure from regional and national firms that automate before smaller firms do. They submit proposals with lower fees and comparable scope. Phase two — likely 12-24 months out — is direct client pressure: clients advised by larger firms or informed by public reporting will start asking smaller CPA firms why their fees haven't reflected AI efficiency gains. The window to automate before clients ask is open now.
Which audit workflows should small CPA firms automate first?
Three categories offer the most immediate return for small audit practices: (1) Transaction classification and sampling — AI tools like Ramp's Accounting Agent achieve 90%+ auto-coding accuracy and significantly reduce time-to-sample. (2) Document classification and intake — AI can sort and categorize support documents before a human auditor reviews them, cutting preliminary fieldwork time. (3) Variance analysis — AI identifies unusual patterns across periods without manual formula-building. These three categories cover most of the non-judgment work in a standard engagement.
Does using AI in audits create any professional liability risk?
The risk is primarily in documentation and oversight, not in the AI use itself. AICPA guidance and most state CPA licensing boards are consistent: AI can assist in audit procedures, but the CPA must verify the AI's work and maintain appropriate documentation. The liability exposure comes when AI outputs are treated as audit evidence without human review — for example, using an AI's transaction classification as-is without testing the AI's accuracy on a sample. Firms that document their AI oversight process are in a strong position. Firms that use AI as a black box are exposed.
Should small CPA firms lower their fees when they use AI?
Not automatically — and not because clients ask. The right framing: AI enables you to deliver the same scope with better accuracy and faster turnaround. That's a quality improvement, not a cost reduction that gets passed through like a supply chain saving. The firms that are succeeding at maintaining rates while using AI are positioning the value as: same coverage, lower error rate, faster close, more time for advisory conversations. They're using AI efficiency to deepen client relationships, not to discount their services.