Tax Departments Can ‘Wrestle the Alligator’ and Use AI Tools Now

A senior tax partner recently shared with me his views on artificial intelligence: “I don’t trust AI and don’t use it.”

He’s not alone. Many tax professionals are steering clear of AI tools in their practice. The work is still getting done. Clients are still being served.

I understand the reluctance. But machines can do so much of what we do faster and at scale. Job disruption is real, and what we’re doing today won’t look the same tomorrow. As a mentor once told me: When you’re wrestling an alligator, it’s better to be on top.

As the vice-president of tax at a global company, I feel the pressure. Management wants to know our AI plan. The team wants reassurance about their livelihoods. Yet I often hear from thought leaders that AI is just a checklist of initiatives to be implemented.

As if it’s that simple. It’s not.

This is my attempt to demystify the moment and share four ways tax professionals can move forward with AI, even if they’re reluctant to do so or are just starting out.

Reframe the anxiety and start experimenting. Many fear AI because they see it competing with us on the same plane: speed and information processing. If that’s the lens, of course it feels threatening.

But that’s the wrong frame. Yes, some roles will change or disappear. That’s already happening. AI isn’t here to replace tax professionals; it’s here to support them. The opportunity isn’t to act more like a machine. It’s to become more human, while letting technology do what it does best.

So start experimenting. Hand your spreadsheets and slides to AI. Iterate. Play with it. Albert Einstein said if he had an hour to solve a problem, he’d spend 55 minutes thinking about the problem and five minutes on the solution. Better questions lead to better answers.

AI is transforming how we learn and work. It accelerates research into any topic: value-added tax in Sweden, stock distribution under Section 355, withholding tax in Saudi Arabia. It expands the universe of content, scanning broader sources than traditional methods allow.

What once took weeks now takes days; what took days now takes hours. And once your research is done, instead of spending evenings summarizing notes, you can generate concise summaries in seconds: analysis to discuss with advisers, short notes for the business, executive ready summaries.

Redesign workflows around objectives, not roles. Traditionally, we design tax functions by defining roles first, then filling them with people. With AI, flip the sequence.

Start with objectives. Then ask which parts of the work are best handled by technology and which truly require human judgment. Real transformation happens when leaders redesign workflows with technology at the center, not when they simply layer tools onto existing structures.

This isn’t about eliminating roles. It’s about redistributing effort, so expertise operates at higher leverage.

Focus on structured, high-volume work. Strong AI use cases come from matching AI’s native strengths (ingesting information, synthesizing it, drafting content, routing workflows) to tax processes that are structured and repeatable.

Transfer pricing is a prime example. Multinationals must maintain a global master file and local reports explaining how the group creates value and why intercompany prices make sense. AI can synthesize functional analyses across entities, harmonize descriptions of risks and assets, and produce consistent narratives for each jurisdiction, anchored in agreements and prior files.

Human judgment still drives the economics, but documentation becomes faster, more consistent, and audit ready.

Research and development credits are another strong fit. AI platforms help determine eligibility, extract data from project records, assemble contemporaneous documentation, and generate audit ready responses, turning what was once a manual process into a continuous, defensible workflow.

State tax apportionment offers similar potential. AI can ingest sales data, apply nexus rules across jurisdictions, flag anomalies in sourcing or allocation, and generate apportionment schedules at scale. Compliance becomes more automated; the team focuses on managing controversy, advising on planning, and ensuring the business understands its exposure.

For most tax departments, building these AI tools internally isn’t realistic. Progress will come from partnering with specialized vendors who can customize their platforms to your company’s specific needs.

The pattern is clear: When AI is aligned with high volume processes, departments gain speed, consistency, and audit readiness.

Preserve learning friction—on purpose. Here’s the paradox: AI removes friction from learning, and that’s both a gift and a risk.

There’s real value in struggling through complex tax work: making mistakes, asking questions, learning by doing. That friction builds judgment. Everyone was once at the moment before they learned the thing they now know.

The solution isn’t to slow people down arbitrarily; it’s to replace passive repetition with active interrogation.

In our tax department, we’ve adapted simulation training to AI outputs. Junior staff present an AI-generated sales tax determination matrix on cloud software to a simulated auditor, then defend the methodology and assumptions. They must explain why the AI made certain choices, where it might be wrong, and what human judgment still matters.

This approach preserves the learning friction that builds judgment while embracing AI’s efficiency gains. The goal isn’t to recreate the old model; it’s to design a new one where questioning AI outputs becomes a core competency.

AI makes answers easier, but growth still requires reflection, feedback, and teaching—now we just need to engineer those moments intentionally.

Looking Ahead

AI in tax is a redesign challenge, not a replacement threat.

The more you feed AI, the better it gets; the same is true for us. Every moment carries everything that came before it, plus something new. I’m 51, but if I’m not constantly upgrading, what’s the point?

The alligator isn’t going away. But if we stay on top and keep upgrading our own version, learning and adapting as the technology evolves, the future of tax can be both more efficient and more human.

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

Jared Dunkin is vice president of tax and senior tax counsel of FTI Consulting Inc., a publicly traded global business advisory firm.

Write for Us: Author Guidelines