How Accounting Firms Are Using AI in 2026: 7 Use Cases With Real Results

Ninety-eight percent of accounting firms use AI every day. Yours is probably one of them — someone on your team is using ChatGPT, Copilot, or a built-in AI feature, whether or not you authorized it and whether or not you know exactly what for.

That's not a guess. That's from Karbon's 2026 State of AI in Accounting Report, which surveyed approximately 600 accounting professionals across six continents. Ninety-eight percent use AI daily. Only 21% have a documented policy.

The problem isn't adoption. The problem is that most firms are getting half the results they could — because they're using AI the way they use a calculator, one-off and individual, instead of the way that actually moves the needle.

This post documents seven ways accounting firms are using AI in 2026 with measurable outcomes. Not hypothetical benefits. Named tools. Specific results. Real workflows that a 5-to-50 person CPA firm can replicate.

1. Tax Preparation and Review

This is where most accounting AI use started, and it's where the clearest ROI lives today.

What firms are doing: AI tools like CCH Axcess Advisor, SurePrep, and TaxDome AI assist with first-draft return review — flagging anomalies, checking that prior-year figures carried forward correctly, identifying deductions that may have been missed, and generating the review comments that senior reviewers used to write manually.

SurePrep's AI reads source documents (W-2s, 1099s, brokerage statements) and populates the return with flagged items for the preparer to confirm or override. The difference between a preparer starting from source documents and a preparer reviewing an AI-assisted draft is significant: it compresses the first-pass review by approximately 30 to 40 percent at firms using the tool consistently.

What small firms can do today: SurePrep and TaxDome AI are accessible to small firms — no enterprise contract required. The entry point is integrating whichever tool connects directly to your existing tax software. CCH Axcess Advisor integrates with CCH Axcess Tax. TaxDome AI connects natively with TaxDome's practice management platform. Before connecting any client tax data to an AI tool, confirm the tool's data processing agreement covers your jurisdiction's client confidentiality requirements.

2. Client Advisory Services (CAS)

CAS is where accounting firms are finding the highest-value AI applications in 2026 — and where the greatest competitive differentiation is opening up.

What firms are doing: Karbon AI and Digits are the most widely deployed tools in this category. Digits connects directly to a client's QuickBooks or Xero and generates a monthly narrative: what changed, why it changed, what the owner needs to know. The accountant's role shifts from preparing that narrative to reviewing and adding judgment.

Karbon AI (built into Karbon's practice management platform) surfaces upcoming client deadlines, drafts status update emails, and flags clients who haven't responded to outstanding requests. For CAS practices specifically, it reduces the overhead of tracking 40-plus advisory clients to something a single staff member can manage.

The ROI case: Thomson Reuters projects a 1:1 ratio of AI agents to accountants by 2027. Firms using Digits-style tools are already operating closer to that model on their CAS client roster — one advisor covering more clients without sacrificing depth.

3. Invoice Review and Accounts Receivable

Most accounting firms manage some version of AR — either for their own firm or as part of bookkeeping services for clients. This is where AI is producing fast, documentable wins.

What firms are doing: Google Workspace Gemini (available in Google Workspace Business Standard and above) can review a batch of invoices for patterns — duplicate charges, vendors with unusual payment timing, categories that spiked month-over-month. QuickBooks AI has embedded anomaly detection and categorization assistance directly in the bookkeeping workflow. CollBox automates client follow-up for outstanding AR by identifying overdue invoices and drafting the collection communication.

What this looks like in practice: A 12-person bookkeeping-heavy accounting firm using QuickBooks AI for categorization review reports roughly 45 minutes saved per client per month on the monthly close — not because AI replaces the review, but because it pre-categorizes and pre-flags, so the reviewer is confirming rather than deciding from scratch.

The AI accounting task automation guide at our archive documents this specific workflow in more detail.

4. Audit and Quality Control

Big 4 headlines about "150 AI agents" don't apply directly to a 10-person CPA firm. But the QC applications do.

What firms are doing: FloQast uses AI to match reconciliations against prior periods, flag variances outside expected ranges, and document the QC checklist automatically as the close progresses. Avalara AVI (Avalara's AI-powered transaction review) identifies sales tax compliance gaps in real time as transactions post. For smaller firms doing attest work, AI-assisted workpaper review catches the procedural gaps that slip through when a reviewer is rushing at deadline.

The IRS angle: GAO data from 2025 shows the IRS scanned 76% more returns for audit indicators using AI compared to prior years. Your clients' returns are being reviewed by AI before a human examiner ever sees them. The practical implication: AI-assisted QC on the preparation side isn't optional for risk management — it's a response to the same technology your clients are now being audited by.

5. Client Communication Drafting

This is the most common daily use of AI across accounting firms in 2026 — and likely where the 98% adoption figure actually lives.

What firms are doing: Staff are using Karbon AI email drafting, Microsoft 365 Copilot in Outlook, and general-purpose tools like ChatGPT to write first-draft client communications. Engagement letter updates, information request follow-ups, deadline reminders, year-end planning outreach — all of these start as an AI draft that a staff member refines and sends.

Karbon AI is the most accounting-specific of these options: it's trained on practice management context, so drafts reference the right client, the right engagement, and the right next action based on the workflow stage. The difference between using Karbon AI for email drafts versus a general-purpose chatbot is that Karbon pulls from the actual client record, reducing the fill-in-the-blanks that generic drafts require.

The documented savings: Karbon's 2026 data shows that firms with a structured approach to AI save 60 minutes per employee per day compared to unstructured use. Email drafting is the highest-volume application — and the gap between firms who've documented which tools to use for which tasks versus those who haven't is most visible here.

6. Research and Regulatory Monitoring

Regulatory complexity in 2026 is significant — and growing. Colorado's AI law took effect June 30. Illinois passed SB 315 (mandatory annual AI safety audits for frontier AI vendors), which means the AI tools your firm uses — Harvey, CoCounsel, Microsoft Copilot — will face annual auditing requirements. Staying current is a real cost.

What firms are doing: Perplexity Pro has become the preferred research tool for time-sensitive regulatory questions at small professional services firms — it surfaces and synthesizes current sources faster than a traditional database search. CCH IntelliConnect and Checkpoint remain the standard for authoritative tax research with proper citation trails.

The shift is workflow: firms that used to assign a staff member to monitor IRS notice releases, state tax updates, and engagement-relevant regulatory changes are now using AI to surface a curated daily brief, with the staff member reviewing for client relevance rather than assembling the research from scratch.

For firms with advisory clients in multiple states: Multi-state regulatory monitoring is where Perplexity Pro or a similar AI research tool pays for itself fastest. The alternative — manually tracking each state's tax authority and AI law developments — is genuinely prohibitive at a 10-person firm.

7. Hiring and Staff Development

This is the use case most accounting firms haven't started yet — and one of the highest-leverage ones available.

What firms are doing: Workday and accounting-specific HR platforms with AI features assist with resume screening and job description drafting. More practically for small firms, AI tools are being used to document standard operating procedures that previously lived only in senior staff members' heads. When a senior accountant can describe how they review a complex return and AI can produce a training document from that conversation, the firm's institutional knowledge becomes less fragile.

The policy connection: Karbon's 2026 data shows that firms with a written AI strategy see approximately twice the adoption rate of firms without one. The mechanism: a policy tells staff which tools are cleared for client work. Without that clarity, staff default to not using AI on anything client-facing, because the risk of using the wrong tool is real and the guidance doesn't exist.

Getting an AI policy in place for your accounting firm is the prerequisite for making any of these seven use cases work consistently. The policy doesn't need to be long — one page covering approved tools, a data handling rule, and who reviews AI output before it reaches a client is sufficient.

What Accounting Firms Are NOT Using AI For Yet

For completeness: these are the areas where AI consistently underperforms expectations at small CPA firms in 2026.

• Judgment-heavy advisory. Tax planning for complex situations — a business sale, a trust restructure, a multi-state acquisition — still requires human judgment that AI can inform but not replace. AI can draft the memo; the analysis has to be yours.
• Novel client situations. When a client's situation doesn't match the training data (new business structure, unusual asset class, unusual filing status), AI tools introduce errors that require experienced review to catch. The risk goes up, not down.
• IRS representation with adversarial dynamics. Correspondence audits with a clear procedural script are manageable with AI assistance. Anything involving negotiation, judgment calls on penalty abatement arguments, or a relationship with an actual examiner still requires a human.
• Complex trust and estate work. The legal and fiduciary complexity, combined with the jurisdiction-specific variations, makes AI-generated output in this area unreliable enough that most experienced practitioners don't use it for substantive analysis.

---

FAQ: How Are Accounting Firms Using AI in 2026?

What percentage of accounting firms are using AI in 2026?

98%, according to Karbon's 2026 State of AI in Accounting Report. The more revealing number is the 77-point policy gap: 79% of those firms use AI daily without any documented guidance on which tools are approved or how client data should be handled.

What AI tools are most commonly used by accounting firms?

The most common tools at small and mid-size CPA firms in 2026 are: Microsoft 365 Copilot (email drafting, document summaries), Karbon AI (practice management, client follow-up), CCH Axcess Advisor (tax prep review), QuickBooks AI (bookkeeping automation), and Google Workspace Gemini (invoice review, research synthesis). The majority of actual daily use is in email drafting and document summaries — not the advanced agentic workflows that get industry coverage.

What accounting tasks is AI actually saving time on?

Karbon's 2026 data shows firms with a structured AI approach save 60 minutes per employee per day compared to unstructured use. The highest-ROI tasks: first-draft tax review commentary, client email drafts, engagement letter generation, and bookkeeping categorization review. The structured-versus-unstructured gap is significant: the same tools, without a documented approach, produce a fraction of the time savings.

Can a small CPA firm (under 10 people) use AI for tax preparation?

Yes. Tools like SurePrep, Canopy, and CCH Axcess Advisor offer AI-assisted tax prep features accessible to small firms without enterprise contracts. The key requirement is setting clear data handling policies before using any AI on client tax data — state AI laws in Colorado (effective June 30, 2026) and Illinois require written AI disclosure frameworks for firms using automated decision tools.

What is the biggest mistake accounting firms make when adopting AI?

Using AI without a written policy. Karbon's 2026 data shows 79% of accounting firms use AI daily without a documented strategy — and those firms save 60 fewer minutes per employee per day than firms with one. The policy doesn't need to be complex: one page covering approved tools, a data handling rule, and a review requirement is sufficient. Without it, staff inconsistency eliminates most of the efficiency gains.

---

The Bottom Line

The seven use cases above cover different parts of a CPA firm's work, but they share a common pattern: AI does the first pass, the human does the judgment call. Tax prep review, CAS narratives, invoice anomaly detection, QC checklists, email drafts, regulatory synthesis, staff training docs — in every case, the value is in the human reviewer moving faster, not in removing the human reviewer.

That's the realistic version of AI for accounting firms in 2026. Not a replacement story. A compression story.

The firms getting the most out of it are the ones who have documented which tools are approved for which tasks — and who have trained their staff to use those tools consistently. The 60-minute gap that Karbon found between firms with a strategy and firms without one is a policy gap, not a technology gap. The tools are available. The question is whether you're using them in a way that compounds or in a way that stays random.

The Crossing Report covers the AI moves that accounting firm owners need to act on — weekly, specific, and filtered for what actually matters to a firm your size. Subscribe here — free, every Monday.