AI Is Gaining Grip on Tax Compliance. Tax Pros Must Lead the Way

Artificial intelligence in tax compliance has shifted from descriptive analytics and anomaly detection toward agentic AI—a system capable of autonomous, multistep decision-making that can work across tax rules, flag exceptions, and drive workflows.

The boundary between humans and machines in compliance operations is blurring. And tax leaders must guide that convergence.

Earlier AI implementations focused on classifying data, mapping general ledger accounts, and spotting anomalies. The current wave, however, emphasizes agentic AI—systems capable of executing multistep tasks with minimal human prompting.

Our “EY.ai for Tax” aims to evolve tax from people executing technology-and-data-powered processes to technology powered by data-executing processes that are managed by people asking questions and reviewing outcomes. These agents can monitor rule changes, propose adjustments to tax provisioning, route exceptions for review, and even draft responses to audit queries.

But their sophistication also raises the bar for integration, data lineage traceability, and human judgment.

Enterprise Tax Functions

The acceleration in tax technology adoption is evident. Several key use cases are emerging now.

Tax notices. AI can read notices received from tax authorities; summarize the key points, amounts and due dates; judge the complexity; draft a response; and log the information in to a database.

Transfer pricing compliance and documentation. AI-powered tools can dynamically evaluate benchmarking and cross-jurisdictional comparables and flag documentation gaps in real time.

R&D credits and incentives. AI is being woven into research and development tax platforms to automate eligibility determination, documentation drafting, and audit response.

Continuous audit and “always-on” monitoring. Instead of batch filings, AI agents analyze compliance metrics continuously, triggering alerts or remediations before deadlines.

Tax research. Large language models can research technical areas and deliver scalable writeups in minutes targeted at different audiences for different purposes.

Tax departments can use AI to shift from reviewing historical transactions to making near-time determinations that drive business decisions. For example, real-time value-added tax calculations enable dynamic pricing, instant duty analysis informs supply chain rerouting, and predictive tax modeling supports mergers and acqusitions negotiations. This transforms tax from a compliance function into a forward-looking partner that influences strategy in the moment it matters.

People, Skills, Culture

Human insight unlocks technology’s full potential. As AI takes over rote compliance tasks, tax professionals must shift to judgment, exception management, model tuning, and strategic insight.

Upskilling in critical thinking, prompt engineering, AI ethics, model validation, and data science is increasingly essential. Tax departments should shift their focus from assembling information to applying critical judgment and conducting high-quality reviews. Professional skepticism is now especially important.

To go back in time when spreadsheets started to become mainstream, reviewers had to understand the inputs as well as the mechanical correctness of the formulas. Today, with AI, we are given conclusions to review, but we can’t just make sure the columns subtotal correctly. We must understand the process and learn how to review in this new world.

There is an increasing divide between those with decades of experience who can identify problems with AI output and those with minimal technical tax experience who may overly rely on the AI-driven results. Both sides should come to the center.

The more experienced generation should learn to leverage the power of AI, while those newer to the profession should harness AI’s capabilities help them improve the clarity of the instructions they provide to AI and build their experience verifying the results.

Questions for 2026

Where in our tax compliance workflows can we deploy agentic AI to reduce cycle times or increase coverage? Evaluate where your people are spending their time reviewing outcomes or performing analytics to test results.

How do we build governance, review, and explainability into all AI systems? AI should be used to improve the human’s review, not to remove the need for the review in the first place. Clear understanding of roles and responsibilities is key. It’s not just whether you used AI; it’s what you did to validate the AI results.

What investments are needed in talent, data infrastructure, and AI tooling to stay competitive? Tax departments should be engaging with others in their organization around these topics as they are likely organization wide. Educating your team and empowering them to look for opportunities to use AI is important. The people most informed by the current process will be vital in identifying areas to implement AI.

What is our cultural path to integrate AI adoption without alienating skeptical team members? Show skeptical team members the power of AI with practical ways to save time and drive better outcomes. Give them an opportunity to explore using the tools and validate the results to gain trust and confidence.

The future of tax isn’t just digital. It’s also intelligent. Agentic systems, empowered professionals, and adaptive governance are converging to redefine what’s possible.

This is a story of elevation, not automation. Tax leaders who embrace this shift will unlock new levels of efficiency, insight, and strategic influence. The frontier is here—and it’s not about doing things faster, but doing them smarter.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Daren R. Campbell is the EY Americas tax technology and transformation leader in Washington, DC.

Randy Carpenter is the EY US business tax compliance transformation leader and serves as a tax account leader in Los Angeles.

The views reflected in this article are the views of the author and do not necessarily reflect the views of Ernst & Young LLP or other members of the global EY organization.

Write for Us: Author Guidelines