We are in the relative early stages of a shift that many tax practitioners still underestimate. Artificial intelligence is already embedded in the systems we use, the data we analyze, and the expectations our clients now bring to the table.
But here is the disconnect.
For the most part, tax planning is still being executed on a timeline built for a pre-AI world. We prepare returns. We reconcile forms. We react to what already happened. Meanwhile, AI is pushing the profession toward something different — a model where planning is continuous, predictive and unavoidable. This is structural. And it is already happening.
The 2026 Thomson Reuters AI in Professional Services Report found that organization-wide AI use in tax and accounting nearly doubled to 40% this year. Most individual professionals are now using generative AI tools. Many firms are already preparing for the next wave: Agentic AI that acts on tasks without human prompting.
Here is what should concern you: Only 18% of firms track ROI on their AI tools. Most corporate clients want their outside firms to use AI on their matters. Less than one-third know whether their firms actually do.
A separate 2026 survey by Accounting Seed found that while 63% of finance teams are exploring AI, only 16% have implemented it in daily workflows. Eighty-four percent of finance teams still spend at least a quarter of their time on manual, repetitive work. The profession is also short more than 200,000 CPA positions. The pipeline is not recovering fast enough. AI is filling the gap that the talent shortage created. And it is doing it faster than most firm leaders expected.
The forcing mechanism we haven't had
Tax season has always created urgency. The IRS demands it. There is a deadline. There are penalties. The system forces compliance. AI is creating a similar forcing mechanism for tax planning. The pressure comes from visibility, not penalties.
When AI can instantly surface missed opportunities, outdated strategies, or inconsistent positions, clients will expect proactive planning the same way they expect accurate returns. The profession has never had this kind of push. Now it does.
And here is the part that should concern every practitioner: The IRS is using AI too. As of early 2026, the agency is deploying machine learning models to score millions of returns simultaneously for audit potential. It is using AI in fraud detection, audit selection, and taxpayer services, in part to offset a workforce reduction of roughly 25%. The agency is doing more with fewer people because AI makes that possible.
If the IRS is using AI to find problems in your clients' returns, you should be using AI to help find them first.
The courts are already punishing careless AI use
If you think the risks of AI are theoretical, look at what is happening in the courts.
In June 2023, a federal judge in the Southern District of New York fined two attorneys $5,000 after they submitted a brief in Mata v. Avianca Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023), containing six entirely fabricated case citations generated by ChatGPT. The cases did not exist. The attorneys did not verify them. That case made national headlines. It should have been a warning.
In September 2025, a California appellate court fined Los Angeles attorney Amir Mostafavi $10,000 after finding that 21 of 23 case quotations in his opening brief were fabricated by ChatGPT. The court published its opinion specifically as a warning to the bar.
On March 16, 2026, the Sixth Circuit imposed more than $100,000 in combined sanctions against two attorneys in Whiting v. City of Athens, 2026 WL 710568 (6th Cir. 2026), after finding more than two dozen fake citations and misrepresentations of fact across three related appeals. The court imposed $15,000 per attorney in punitive sanctions, ordered reimbursement of the opposing party's attorneys' fees, and awarded double costs. The court wrote that the attorneys had "sullied the reputation of our bar."
The pattern is clear. The fines are getting larger. The consequences are getting more severe. Courts are treating this as a breach of professional duty.
For tax professionals, the lesson is direct. AI will generate plausible-sounding authorities that do not exist. It will cite regulations, rulings, and cases that look real but are fabricated. If you use AI to draft a protest letter, a Tax Court petition, or a technical memorandum and you do not verify every citation, you own the consequences.
The liability question is two-sided
The Pennsylvania CPA Journal published a piece in late 2025 on AI and professional liability. The core message: AI creates malpractice risk on both sides.
Use AI carelessly, or rely on its output without verification, and you own the errors it produces. Ignore AI entirely, and you risk falling below the emerging standard of care. When available tools could have identified a planning opportunity or a compliance risk, and you did not use them, that gap becomes a liability question. The standard of care moves with the profession. AI is moving it now.
Compliance and planning are merging
Many practitioners are missing this. AI does not distinguish between compliance data and planning data. It ingests everything.
A change in ownership structure triggers planning implications. A shift in cash flow triggers entity-level modeling. A new jurisdiction triggers cross-border analysis. A missed election triggers a risk alert.
The return is no longer the end of the process. It is the data feed that powers the next decision. AI is replacing the version of the profession that waits for documents, reacts to events, and treats planning as a once-a-year conversation. Our job has shifted from processing information to interpreting it. From reacting to advising.
The firms that will make it through this transition will build AI-driven planning into their workflow. They will train staff to interpret AI outputs, not fear them. They will verify every citation, every authority, every conclusion the software produces. They will integrate tax, financial, and entity-level data streams. They will move from episodic planning to continuous strategy.
The firms that do not will fall behind. Not gradually. Quickly.
Clients assume we are looking at the whole picture. AI will make that assumption explicit.
If we are not using these tools, we will miss what the software sees instantly. If we are not raising issues, AI will. And clients will wonder why we did not.
The IRS is already using AI to find what we missed. The courts are already punishing professionals who use AI without verification. Our clients will soon use AI to check our work.
The question is not whether AI will reshape tax planning. It already has.
The tax return gets filed every year. The planning opportunities surface every day.
