What did the Caseware/IDC study find about AI in accounting firms?

A global study commissioned by Caseware and conducted by IDC surveyed 1,005 audit and accounting professionals across eight regions. The study found that 66% of firms have AI embedded in their strategy — either piloting, deploying in select functions, or running AI at scale. However, 55% of respondents said they would trade AI performance for stronger security or safety controls, and 66% called for a globally harmonized AI framework for audit and assurance. The top adoption barriers were cost of implementation (34%) and lack of technical talent (30%). The study was released in March 2026.

What are the three stages of AI adoption in accounting firms?

The Caseware/IDC study describes three tiers: Experimenting (piloting AI in isolated use cases, no firm-wide integration), Expanding (deploying AI in select practice areas or workflow functions with some systematic adoption), and Embedding (AI is integrated into core strategy and multiple practice areas with governance in place). Most small accounting firms are in the Experimenting tier — which the study notes is exactly where the cost and talent barriers stall adoption. The gap between Experimenting and Expanding is the most common stuck point.

Why are small accounting firms stuck in pilot mode?

The research, supported by CPA Practice Advisor's March 2026 analysis, identifies the root cause: firms are stuck in pilot mode because one person tried a tool, the workflow wasn't shared firm-wide, there are no standard procedures, and there is no governance framework. The result is individual AI use that doesn't compound into firm capability. The technical problem is usually not the AI tool — it's the absence of a shared workflow and a clear internal policy on approved tools and use cases.

What does the call for global AI standards mean for small accounting firms?

66% of accounting professionals surveyed are calling for globally harmonized AI standards for audit and assurance. For a small firm, this signals that external governance requirements are likely to arrive in the next 12-24 months — whether from regulators, the AICPA, international standards bodies, or state bar equivalents. Firms that build internal AI governance frameworks now will have a significantly easier time complying with external requirements when they arrive. The firms that wait for the standard to be handed to them will face a more expensive and disruptive compliance exercise.

What is the minimum viable AI governance framework for a small accounting firm?

Four components: (1) An approved tools list — which AI tools are permitted for client work, internal work, and which are not permitted at all. (2) An acceptable use policy — what can AI be used for and what requires human review before client delivery. (3) Client data handling rules — which client data can be processed through AI tools and which cannot (particularly important for PII, financial data, and attorney-client equivalents). (4) Output review requirement — all AI-generated client-facing work must be reviewed by a licensed professional before delivery. This is less than one day of work to document and is sufficient to satisfy most current and anticipated governance requirements.

Caseware IDC 2026: AI in Accounting Firms

Published: March 14, 2026 | By: The Crossing Report | 6 min read

Summary

A global study commissioned by Caseware and conducted by IDC found that 66% of accounting firms now have AI embedded in their strategy — but fewer than 1 in 5 small organizations meet the readiness thresholds for talent, technology, or governance. Two-thirds of accounting professionals are calling for globally harmonized AI standards before they go further. Here is the three-tier self-assessment, what the profession-wide data says about small firms specifically, and the minimum governance framework a 5-25 person accounting firm can build in a day.

The Gap Between "Using AI" and "Ready for AI"

The headline number from the Caseware/IDC study is encouraging on its face: 66% of accounting firms have AI embedded in their strategy. But the next layer of data is where the real picture emerges.

Only 24–27% of organizations in the study report adequate AI talent, IT readiness, or regulatory preparedness. Among smaller organizations specifically, fewer than 1 in 5 meet any of those thresholds. The gap between "we have an AI strategy" and "we are ready to execute that strategy" is wide — and it is widest at smaller firms.

This is the stuck-in-pilot-mode problem that CPA Practice Advisor's March 2026 analysis identified as the root cause of adoption failure in small accounting firms: one person tried a tool, the workflow wasn't shared firm-wide, there are no standard procedures, and there is no governance framework. Individual AI use doesn't compound into firm capability without those pieces.

The encouraging counterpoint in the same data: firms that have crossed the threshold — early adopters with embedded AI strategies and governance in place — report a 73% rate of strategic advantage. The gap between the stuck majority and the compounding minority is not talent or tools. It is workflow standardization and governance.

Three Tiers of AI Adoption in Accounting

Based on the Caseware/IDC framework and the CPA Practice Advisor analysis, most accounting firms fall into one of three tiers.

Tier 1 — Experimenting You or someone at your firm has used AI tools — ChatGPT for client emails, tax research via AI-assisted search, meeting transcription. There is no firm-wide deployment, no approved tools list, no standard workflow. AI use is individual and inconsistent. Most firms under 15 people are here.

The characteristic feeling at this tier: AI is interesting but the ROI isn't clear. Each person is making their own decisions about which tools to use and for what, which means the firm isn't accumulating institutional capability — just individual habits.

Tier 2 — Expanding You have deployed AI in at least one consistent workflow that involves more than one person at the firm. There is an implicit understanding of which tools are approved and for what. You have started documenting AI-assisted outputs differently from fully manual work. You're seeing time savings in specific areas but haven't quantified them.

The characteristic feeling at this tier: AI is working in some areas, but you're not sure how to roll it out more systematically or what the governance requirements are.

Tier 3 — Embedding AI is integrated into multiple practice area workflows with governance in place. You have a written policy on approved tools, acceptable use, and client data handling. You have quantified time savings in at least two workflow areas. You are evaluating AI tools against a criteria framework rather than adopting them ad hoc.

The characteristic feeling at this tier: AI is a routine part of operations that your team talks about, measures, and improves.

The Global Standards Signal

The most significant forward-looking finding in the Caseware/IDC study: 66% of accounting professionals are calling for globally harmonized AI standards for audit and assurance before they go further with deployment.

For a small firm owner, this number is a signal about the compliance landscape ahead.

The profession's leading bodies — AICPA, IFAC, PCAOB, and their international equivalents — are watching the same data and hearing the same demand. Globally harmonized AI standards for accounting and audit will arrive. The timeline is uncertain, but the direction is not. When those standards land, they will require documentation of your AI governance practices, a log of AI tools used in client-facing work, and evidence of human review protocols.

Firms that build internal governance frameworks now — even lightweight ones — will adapt to external standards in hours, not weeks. Firms that haven't documented their AI practices at all will face a more disruptive compliance exercise when the standard arrives.

The Minimum Viable Governance Framework

Here is what "AI governance" looks like for a 5-25 person accounting firm — not the enterprise version, the version a firm owner can build in one afternoon.

1. Approved tools list (30 minutes) Write down every AI tool currently in use at your firm. For each one, categorize it: approved for client-facing work, approved for internal work only, or not approved. The list does not need to be long. It needs to exist.

2. Acceptable use policy (45 minutes) One-page document. What can AI be used for at your firm? What requires human review before it goes to a client? What cannot be processed through AI tools (typically: unencrypted client financial data, PII, tax ID numbers in unapproved platforms)? This document is both an internal policy and a starting point for responding to future regulatory requirements.

3. Client data handling rules (20 minutes) Identify which AI tools in your approved list process client data, and confirm each tool's data handling practices (are client inputs used to train the model? is data stored? in what jurisdiction?). Tools that process client data must be evaluated against your engagement letters and any applicable data protection obligations.

4. Output review requirement (immediate) Establish the default: all AI-generated content that goes to a client must be reviewed and signed off by a licensed professional before delivery. Document this in your acceptable use policy. This is the single most important governance step for managing professional liability exposure from AI use.

These four components take less than two hours to document. They are sufficient to satisfy most current and anticipated governance requirements — and they move your firm from Tier 1 to the bottom of Tier 2 in a single afternoon.

Where Small Firms Actually Have an Advantage

There is a framing problem in how the accounting profession talks about AI adoption: the enterprise frame. Large firm AI coverage is about multi-million dollar implementations, firmwide rollouts, and governance committees. That frame makes small firm owners feel behind.

The Caseware/IDC data tells a different story. The adoption barriers for small firms — cost of implementation (34%) and lack of technical talent (30%) — are real, but they are not as prohibitive as the enterprise framing suggests. The AI tools that drive measurable efficiency for a 10-person accounting firm are not the same tools that a Big 4 firm deploys. They are ChatGPT Plus ($20/month), Black Ore Tax Autopilot ($99-$299/month), and Karbon with AI meeting summaries (already in your subscription). The governance framework above doesn't require a compliance department.

The early adopters generating 73% strategic advantage in the Caseware/IDC data are not all large firms. They are firms that standardized one workflow, documented it, and compounded from there. That path is as available to a 12-person CPA practice as it is to a regional firm.

The question is not whether you have the resources. It is whether you build the habit before external standards arrive to require it.

66% of Accounting Firms Are Using AI — Where Does Yours Stand?