The Complete Guide to AI for Tax Professionals in 2026: Tools, Time Savings, and Compliance Rules

AI adoption among tax professionals has hit an inflection point. In 2024, just 9% of tax and accounting firms had formally adopted AI. By 2025, that number was 41%. By early 2026, 46% of accountants report using AI daily — up from 18% just three years ago. AI is no longer a curiosity in tax practice. It is infrastructure.

The question is no longer whether to use AI. The question is which tools to use, what those tools can actually do, where professional standards draw the line, and how to use AI in a way that does not expose you or your clients to liability. This guide covers all of it.

The Current State of AI Adoption in Tax Practice

The Wolters Kluwer Future Ready Accountant 2025 report found that formal AI adoption in tax and accounting firms jumped from 9% in 2024 to 41% in 2025 — a 4.5x increase in a single year.¹ CPA.com's 2025 AI in Accounting Report found that 58% of accounting firms are piloting AI, and 70% of US tax professionals have adopted some form of AI-driven process.²

The driver is time. Tax professionals are overwhelmed by volume: research that once took three to five hours now takes 15 to 30 minutes with AI assistance. Early testing of Thomson Reuters' Ready to Review platform found that users reduced time spent on a simple 1040 return by approximately one hour — and the average CPA handles hundreds of returns per season.³

The productivity case is real. But productivity without compliance is a liability. The same survey data that shows surging AI adoption shows a critical gap: most professionals are using AI without a formal policy, without proper contracts, and without a clear understanding of what their professional standards require.

The Major AI Tools Being Used in Tax Practice in 2026

The market for tax-specific AI has matured significantly. There are now purpose-built tools designed for tax professionals that offer enterprise-grade data protections — a very different risk profile from consumer ChatGPT.

Thomson Reuters CoCounsel & Ready to Review

Thomson Reuters launched its fully deployed agentic AI platform in March 2025 and now has thousands of firms using it. The flagship product for tax preparation is Ready to Review: it automates US 1040 Form preparation by processing source documents and prior-year returns. Early testers reduced preparation time by approximately one hour per simple return. Ready to Review reached general availability in December 2025. Thomson Reuters also launched Ready to Advise for tax advisory workflows, and CoCounsel Document Analysis (January 2026) for complex audit workflows.³

Bloomberg Tax AI Assistant

The Bloomberg Tax AI Assistant won the 2025 CPA Practice Advisor Technology Innovation Award. It is designed for tax research: complex questions in natural language, with grounded answers backed by Bloomberg's primary source database. A recent Bloomberg survey found that 84% of tax professionals using the tool reported finding insights faster, reducing hours of work into minutes.⁵

Black Ore Tax Autopilot

Black Ore launched Tax Autopilot for broad availability in April 2026. It targets enterprise-grade AI for the tax profession, with full data security controls and a focus on the firm-level workflow, not just individual research tasks.⁶

Juno

Juno is an AI tax preparation platform that raised $12 million in seed funding in April 2026. It claims to automate 90% of busy work, with firms reporting 50% less time spent per return.⁷

General-Purpose AI (ChatGPT, Claude, Gemini)

Many CPAs use general-purpose AI tools for tax research, drafting client communications, and understanding complex documents. These tools can be highly effective for research that contains no client-specific information. The compliance issues arise when professionals input client data — which we cover in detail below.

What AI Can and Cannot Do for Tax Work

AI in tax practice is enormously capable in specific areas and genuinely unreliable in others. Understanding this distinction is the difference between using AI as a powerful tool and using it as a liability generator.

Where AI genuinely helps

• Document comprehension. AI excels at explaining complex language in plain terms. A confusing clause in a partnership agreement, an ambiguous line on an IRS notice, or a technical section of the Internal Revenue Code — AI can explain these clearly and quickly.
• Research acceleration. Finding relevant code sections, rulings, and guidance used to take hours. With purpose-built tax AI tools, it takes minutes. Bloomberg Tax AI Assistant showed 84% of users finding insights faster.⁵
• Data entry and form population. AI automation reduces manual data-entry errors by 75% in firms that have implemented it — a dramatic improvement in return accuracy for routine filings.⁴
• Drafting client communications. AI can draft clear, readable client letters that explain complex tax matters in plain language. This saves time and often improves quality.
• Consistency checking. AI tools can cross-check return line items against code requirements and flag potential issues before filing — reducing audit risk by approximately 40% for firms using sophisticated AI error-detection.⁸

Where AI fails — and the risks are severe

• Hallucination. AI confidently generates responses that are wrong. In the 2023 Mata v. Avianca case, attorneys relying on ChatGPT-generated research cited fictitious cases — earning judicial sanctions. By late 2025, nearly 800 documented cases of AI-related citation errors or hallucinations had been logged across at least 25 countries.⁹
• Current law. AI models have knowledge cutoffs and may not reflect the most recent legislative changes, IRS guidance, or court decisions. Always verify research against current primary sources.
• Tax strategy and positioning. AI is a comprehension and research tool, not a strategic advisor. Optimizing a client's tax position requires judgment, experience, and an understanding of the client's full situation that AI cannot replicate from a single prompt.
• Ambiguous or complex transactions. AI struggles with situations involving genuine interpretive ambiguity — exactly the situations where experienced professional judgment matters most.

The Compliance Minefield: What Your Professional Standards Actually Require

The professional standards governing AI use in tax practice are more specific than most CPAs realize — and the consequences of getting them wrong range from disciplinary action to criminal liability.

AICPA Statement on Standards for Tax Services (SSTS)

SSTS Section 1.4.2 defines a “tool” as including “artificial intelligence.” Tax professionals retain all professional obligations whether or not they use a generative AI system. The AICPA's guidance is unambiguous: AI is a tool, like a calculator or a tax preparation platform. The professional responsibility for accuracy and compliance sits with the CPA, not the tool.¹⁰

IRS Circular 230

Circular 230 governs practice before the IRS and imposes duties of competence, diligence, and confidentiality. Using consumer ChatGPT to process client tax data may violate the confidentiality obligations embedded in Circular 230. The IRS formalized its AI governance framework in IRS Information Release 10.24.1, issued in March 2025, and Treasury published formal AI user agreement guidance in October 2025.¹¹

IRC §7216 — The Criminal Statute

This is the provision most CPAs do not think about when they paste a client's tax return into ChatGPT. IRC §7216 makes it a federal crime to knowingly or recklessly disclose information furnished in connection with a tax return. Violations carry penalties up to $1,000 per violation and up to one year in prison.

When you input a client's tax return information into a consumer AI platform, that information is transmitted to and processed by OpenAI (or Google, or Microsoft) — a third party with no authorization to receive it. The “knowing” standard may well be met if you are aware of what these platforms do with your data and submit client information anyway.¹²

The Heppner Ruling and What It Means for Tax Advisors

United States v. Heppner (S.D.N.Y., February 17, 2026) established that public AI tool conversations are not protected by privilege. The court found that by submitting legal strategy to ChatGPT, the defendant had voluntarily disclosed privileged information to a third party, waiving the privilege.

Morgan Lewis analyzed the tax implications in March 2026: “Tax advisors who input client information into public AI tools should understand that those interactions are not shielded by any professional privilege. If that information becomes relevant to litigation or an IRS audit, it may be discoverable.”¹³

Proskauer Rose went further, noting that tax advisor arrangements like Kovel agreements — designed to extend attorney-client privilege to CPA work — may be undermined when privileged tax strategy is submitted to an AI platform that has no confidentiality obligation.¹⁴

The practical implication: any tax strategy memo, any analysis of a client's exposure, any documentation of a contentious filing position — if drafted or researched using consumer AI, that privileged work product may no longer be privileged.

A Framework for Safe AI Use in Tax Practice

Given the compliance requirements and the real risks, here is a practical framework for using AI in tax practice without creating liability:

Tier 1: Public AI Tools (ChatGPT free, Gemini, Bing Copilot)

Permitted use: Research that contains zero client-specific information. Understanding a general code provision. Drafting a template letter with no client data. Learning about a new area of tax law.

Prohibited use: Anything involving a client's name, Social Security number, financial details, business information, legal position, or any other information that could identify the client or constitute a disclosure under IRC §7216.

Tier 2: Enterprise AI with Data Protections (ChatGPT Enterprise, specialized tax AI)

Permitted use: Research and drafting with client data, provided your engagement agreements and privacy policies reflect the use, you have a valid contract with the vendor including zero-retention clauses, and you have verified the vendor's compliance with applicable data protection requirements.

Required due diligence: Review vendor contracts carefully. “We do not train on your data” is different from “We do not retain your data.” Both protections are required. Verify that the contract addresses IRC §7216 compliance.

Tier 3: On-Device AI

Permitted use: Any client work, without exception. Because the AI model runs entirely on your hardware, data never leaves your device. There is no transmission to a third party, no retention by a vendor, and no IRC §7216 disclosure.

Compliance position: On-device AI eliminates the entire category of third-party disclosure risk. For the most sensitive client work — audit defense, tax litigation support, strategic planning involving confidential positions — on-device AI is the only option that is clean under every applicable standard.

The IRS Is Using AI Too — Know What That Means

In late 2025, the IRS deployed Agentforce AI across the Office of Chief Counsel, Taxpayer Advocate Services, and the Office of Appeals. The system searches documents and provides case summaries, accelerating the IRS's ability to identify issues and process cases.¹¹ The IRS is also exploring using AI to identify credits and deductions taxpayers may have missed — and, by extension, to identify compliance patterns that may flag returns for review.

The GAO has called for better documentation and transparency around IRS AI use, and advisory panels have recommended improvements in how the IRS communicates its AI-assisted decisions to taxpayers.¹¹ But the trajectory is clear: the counterparty in a tax dispute is increasingly AI-assisted. Your practice should be too.

Taxpayer Trust Is Shifting Back to Human CPAs — Use That

Consumer trust in AI for financial matters is declining. A 2026 survey found that 43% of taxpayers trusted AI for tax filing in 2025 — but that number dropped to 37% in 2026 as taxpayers shifted back toward human tax professionals. Only 19% of Americans trust AI in financial services, and only 10% trust AI to make financial decisions automatically.¹⁵

This is an opportunity for CPAs who use AI correctly. Clients want the combination: the speed and comprehensiveness that AI enables, delivered by a human professional they trust, using tools that are designed not to expose their data. The CPA who can say “I use AI to work faster and more accurately, and the tools I use never share your data with anyone” has a genuine competitive advantage over one who is either avoiding AI entirely or using it unsafely.

AI in tax practice is not coming — it is here. The firms that will lead in 2026 and beyond are not those that adopt AI most aggressively. They are those that adopt it most intelligently: understanding what the tools can and cannot do, building clear policies that protect client confidentiality and professional privilege, and choosing tools whose architecture — not just their contracts — guarantees that client data stays where it belongs.

The technology is ready. The question is whether your practice is.