AI Agents for Corporate Tax Departments: Why 64% Are Still Stuck in Reactive Mode

Corporate tax teams have been told for a decade that technology would free them up for strategic work. For some, it has. For most, the latest data says the opposite is happening.

The 2026 Corporate Tax Department Technology Report from the Thomson Reuters Institute and Tax Executives Institute surveyed 170 tax leaders at companies of all sizes. Its central finding is what the authors call a "frustration gap": tax professionals know what better tools could do for them, but budgets, purchasing and training aren't keeping up. And the share of departments still running in firefighting mode is going up, not down.

Most Departments Are Still Reactive

The report places departments on a five-step maturity curve, from chaotic (email and spreadsheets) through reactive, proactive and optimized, up to predictive. According to Thomson Reuters' summary of the findings:

• 64% of departments describe themselves as chaotic or reactive, up from 57% the year before
• The proactive group shrank from 35% to 29%
• Only 5% have reached optimized, and 2% predictive

That is a striking result after years of investment. The work still gets done, but often through a heroic effort at provision time and a compliance season that is survived rather than run.

The Technology Is Helping the Teams That Have It

The report isn't all bad news. Two-thirds (67%) of respondents say technology investment over the past three years has helped their department shift toward more strategic, proactive work, with 23% calling the shift significant. The effect is strongest in larger departments.

Leadership has also caught up. In the latest survey, 88% of departments have a dedicated person responsible for tax technology strategy, up from 51% the year before. In two-thirds of cases, that person is the head of tax.

So the problem isn't that tax leaders don't care or don't know what to do. It's that the tools and money aren't there for most of them.

Budgets, Purchases and Satisfaction Are All Moving the Wrong Way

Three numbers from the report explain the frustration:

• About half of departments bought no new tax technology at all in the past 12 months.
• 56% are not satisfied with their tech stack, up from 34% the year before. The "very satisfied" share fell from 11% to 4%.
• 69% expect their technology budget to rise next year, but only about 39% of last year's respondents actually got the increase they expected. Just 27% say they can secure the budget and resources they need when they need them.

Training is slipping too. The share of companies offering ongoing tech training fell from 59% to 50%.

Tax Leaders Now Expect AI Within Two Years

The biggest change in the report is how quickly tax professionals expect AI to arrive. A year earlier, the most common answer to "when will AI be central to your workflow?" was three to five years. Now it is one to two years (39%). The share saying AI is already central rose from 2% to 7%.

When asked where AI would add the most value, respondents ranked:

1. Automating routine compliance (69%)
2. Accelerating technical research (61%)
3. Error detection (36%)
4. Generating insights from tax data (31%)
5. Document review (31%)
6. Scenario modeling (25%)

Put the two halves of the report together and the picture is clear. Tax teams want AI for routine work, they expect it soon, and many of them won't get a big new software budget to pay for it.

Sign up for our newsletter

Get the latest updates, product launches, and helpful tips delivered straight to your inbox.

Email addressSubscribe

Where AI Agents Fit in a Tax Department

That combination is exactly where AI agents can help. Instead of replacing the tax engine or the ERP, an agent works across the systems and spreadsheets the team already has, doing the gathering, checking and drafting that eats up the calendar.

Data collection for the provision and returns. Much of a tax team's close is chasing numbers: trial balances, fixed asset rollforwards, intercompany schedules and business-unit questionnaires. An agent can send the requests, track what's missing, pull the files into a standard format and flag anything that doesn't tie out.

Notice triage. State and local notices arrive by mail and email in every format. An agent can log each one, identify the jurisdiction, tax type, period and response deadline, match it to the relevant return and route it to the right person with a draft response outline.

Technical research support. Respondents ranked faster research second on their list. An agent can monitor rule changes in the jurisdictions the company files in, summarize what changed with links to the primary source, and note which entities might be affected. A tax professional still decides what it means.

Error detection before filing. Agents can compare current-year workpapers against prior years and flag unusual swings, missing schedules or apportionment factors that moved more than expected, so reviewers spend their time on the items that matter.

Documentation and audit support. Agents can assemble the support for a position or an audit request from the files already on hand and keep an index of what was sent and when.

Keeping Judgment Where It Belongs

Tax positions carry real penalties and real exposure. AI agents should prepare, not decide. The right pattern is an agent that drafts, reconciles and flags, with a named reviewer who approves anything that goes to a tax authority or into a filed return. Every output should link back to its source so the reviewer can check it quickly.

That also helps with a problem the report raises directly: two-thirds of departments that measure technology performance track time savings, but only 41% track ROI and cost savings. An agent that logs what it did, and how long the same task used to take, makes the ROI case much easier to build for finance and IT.

What This Means for Tax Departments Going Forward

The Thomson Reuters data describes a profession that is ready for AI but stuck between rising expectations and flat budgets. Waiting for a large platform purchase may mean waiting a long time, since half of departments bought nothing new last year.

The departments that pull ahead will likely be the ones that start with the routine work they already understand, such as data requests, notices, reconciliations and research summaries, and automate it in steps with a person reviewing every output. That is how a reactive team becomes a proactive one without a big-bang project.

Frequently Asked Questions

Who conducted the 2026 Corporate Tax Department Technology Report? The Thomson Reuters Institute, together with the Tax Executives Institute. It surveyed 170 corporate tax leaders from companies of all sizes.

What is the "frustration gap"? It's the report's term for the gap between what tax professionals know better technology could do and what their current tools, budgets and resources actually allow.

What do tax teams most want AI to do? Automate routine compliance (69%) and speed up technical research (61%), followed by error detection, data insights, document review and scenario modeling.

Should an AI agent file returns or respond to notices on its own? No. Agents are best used to gather data, draft responses and flag issues, with a qualified tax professional reviewing and approving anything that is filed or sent.