LucaLedgerLucaLedger

Home › Resources › Guides › AI in Accounting 2026

How AI is changing accounting firms in 2026

By Jonathan Solomon, Founder of LucaLedger. Last updated June 2026.

The short answer Enterprise GenAI use among tax and accounting firms roughly tripled in a year, from 8% in 2024 to 21% in 2025 (Thomson Reuters Institute, 2025), and 79% of firms expect significant integration by 2027. But adoption is outrunning depth: only 13% say AI is central to their workflow today. The real 2026 story is the widening gap between firms that have operationalized AI and those still experimenting, and the four forces driving it: tripling adoption, the shift from AI that assists to AI that does the work, a governance bottleneck, and a talent crunch that makes automation non-optional.
Schedule a demo

1. Adoption tripled, but depth is lagging

The headline number is striking: enterprise-level GenAI use among tax, audit, and accounting firms jumped from 8% in 2024 to 21% in 2025 (Thomson Reuters Institute, 2025 Generative AI in Professional Services Report). Sentiment is nearly universal, 89% of tax-firm professionals believe GenAI can apply to their work, and 71% say it should (up from 52% the year before).

Yet capability and practice have separated. Only 13% of tax-firm professionals say GenAI is currently a central part of their workflow (Thomson Reuters Institute, 2025). The AICPA & CIMA's February 2026 global survey frames the same divide differently: only about a quarter of organizations qualify as “AI-Transformed,” and North America sits below the global curve on strategic AI impact.

What it means: in 2026, the question isn't whether a firm uses AI, most have touched it. It's whether they've moved past dabbling. The firms pulling ahead aren't the ones with the most tools; they're the ones that have rebuilt a workflow around AI and trust the output enough to rely on it.

2. The agentic leap: from “assist” to “do the work”

2025 was the year the major platforms shipped agents, not chatbots. Thomson Reuters launched “Ready to Review,” which drafts a 1040 from source documents and prior-year returns and resolves diagnostics (GA December 2025). Wolters Kluwer launched CCH Axcess Expert AI (October 2025) for agentic research, extraction, and audit. Intuit rolled AI agents into its accountant platform across 2025.

This changes the competitive question. Through 2024, the differentiator was “does it have AI?” In 2026 it's “what can the AI actually execute end-to-end, and can you trust it?” The most useful AI in an accounting firm isn't a chat window bolted onto existing software, it's AI woven through the work itself: reading the document, keeping the books, drafting the return, with the accountant reviewing exceptions.

Top reported GenAI use cases reflect this shift toward core work: tax research (77%), tax return preparation (63%), tax advisory (62%), and bookkeeping (57%) (Thomson Reuters Institute, 2025). The frontier is connecting those steps so data flows between them instead of being re-keyed.

3. Governance is the bottleneck, not capability

The most revealing 2026 numbers aren't about technology. They're about readiness. Among tax firms, 70% have no GenAI usage policy, 72% offer no GenAI training, and only 20% measure AI's ROI (Thomson Reuters Institute, 2025). The AICPA & CIMA found only about a quarter of organizations report adequate AI talent, IT readiness, or regulatory preparedness (2026).

And the gap is expensive to leave open, because training demonstrably pays: firms that invest in AI training report saving meaningfully more time per employee (Karbon, State of AI in Accounting, 2025, note: self-selected, AI-leaning sample, so treat as a ceiling). Data security is the dominant concern, and it's rising as firms handle more client data through AI tools.

What it means: the firms that win with AI in 2026 are the ones that treat governance, a written AI policy, staff training, security controls, client disclosure, and a human review step, as the enabler, not the paperwork. For client tax data, “move fast” without “and verify” isn't a strategy; it's a liability.

4. The talent crunch makes automation non-optional

Underneath the AI story is a demographic one. The accounting talent pipeline is shrinking, the number of people sitting for the CPA exam has fallen sharply over the past decade, while the Bureau of Labor Statistics projects roughly 136,000 accounting and auditing openings a year through the mid-2030s (BLS; AICPA/NASBA Trends data).

Firms can't hire their way out the way they once could. So AI adoption in 2026 is less an efficiency upgrade and more a capacity-survival strategy, a way to serve more clients with the staff a firm can actually retain. At the same time, automating the compliance grind is pushing firms up-market toward advisory, where margins are higher and the human judgment is the product.

Even the regulator is moving: the IRS reported 126 active AI use cases as of mid-2025, including machine-learning models for audit selection (U.S. GAO, 2025 to 2026). AI isn't coming to the profession, it's already on both sides of the desk.

What this means for your firm

Where LucaLedger fits

LucaLedger is built for the agentic, all-in-one shift this data describes: one AI-native platform where the document is read once and flows through AI bookkeeping, tax preparation (1040/1120/1120-S/1065 across 46 states + DC), and planning, with the accountant reviewing exceptions, not re-keying data between tools. On the governance point (trend 3), security is foundational: LucaLedger is SOC 2 ready, with encryption in transit and at rest and row-level isolation, and every line on a return is traceable to its source. E-filing closes the loop in November 2026.

And on the agentic shift (trend 2) specifically: LucaLedger is piloting its own firm-scoped AI agent, you talk to it to pull up a client's documents, books, financials, or return and act across them, with every change proposed for you to confirm before it's applied. Because it works inside the all-in-one platform, it spans the whole engagement, not one tool's slice. (In limited pilot, rolling out at general availability.)

We're not claiming AI replaces the accountant, the data says the opposite, that judgment and review matter more as automation grows. We're building the platform that does the repetitive work so the firm's people can do the rest. See how it works, or schedule a demo.

Frequently asked questions

What percentage of accounting firms use AI in 2026?

Enterprise-level GenAI use among tax and accounting firms reached 21% in 2025, up from 8% in 2024, and 79% of firms expect significant integration by 2027 (Thomson Reuters Institute, 2025). Broader “automation” use is far higher; deep, central AI integration is still the minority (about 13%).

What do accounting firms use AI for?

The top uses are tax research, tax return preparation, tax advisory, and bookkeeping, followed by document summarization and review (Thomson Reuters Institute, 2025). Adoption is fastest where AI touches core, repetitive work.

Is AI going to replace accountants?

The data points the other way: as automation grows, human review, judgment, and advisory become more valuable, and most firms report AI shifting their people toward higher-value work rather than eliminating roles. The talent shortage is a bigger pressure than displacement.

What's the biggest barrier to AI in accounting firms?

Governance and readiness, not technology, 70% of tax firms have no AI policy and 72% offer no training (Thomson Reuters Institute, 2025), and data security is the top concern. Firms that invest in policy and training get more out of AI.

Is AI safe to use with client tax data?

Only with the right controls, security, a human-review step, and a clear policy. Platforms handling tax data should be able to show their security posture; LucaLedger is SOC 2 ready, with encryption and row-level isolation.

Schedule a demo

Sources