We are reaching an important turning point in how AI changes work.
For the past several years, conversations about AI have focused on what generative AI can produce: finding information, analyzing data, writing code, drafting content. Agentic AI introduces a different possibility.
Instead of only responding to a request, AI can now understand context, reason across information, coordinate multiple steps, and act within defined parameters. That changes the question from “How can AI help us do this task faster?” to “How should this work be done in the first place?”
I believe that distinction will reshape entire industries, and tax and compliance may be one of the clearest examples.
Why Tax and Compliance Is the Proving Ground for AI Agents
Tax and compliance is extraordinarily complex by nature. Businesses operate across thousands of jurisdictions, sell through expanding channels, launch new products, and respond to regulations that never stop changing. At the same time, the expectation for accuracy is absolute. Technology has automated pieces of that complexity for decades. The opportunity now is much bigger.
Traditionally, compliance has followed business activities.
A company enters a market, launches a product, or absorbs a regulatory change. Someone needs to notice it, understand it, and update the right systems. That model gets harder every year as a business grows.
Imagine instead a compliance system that continuously understands what is changing, inside your business and across the regulatory landscape. When a new product, market, or rule creates a compliance implication, an agent can flag it, recommend an action, and, within defined controls, help execute it.
The expert human remains in command, but no longer must manually find every change. That moves compliance from reactive to continuous, which matters beyond the tax department. It shapes decisions about where to sell, what to sell, and how fast a company can enter a new market. The better a business can manage those obligations, the more confidently it can grow.
Trust Comes First
Tax and compliance is a high-stakes environment. An answer that is right most of the time isn’t good enough when it determines how much tax a business collects, what it reports to a government, or whether it is meeting its regulatory obligations.
That means the future of agentic tax and compliance cannot be built on AI alone.
It requires trusted data, deep domain expertise, precise calculation, clear controls, transparency, and auditability. Businesses need to know not only what an intelligent system recommends or does, but why.
This is where the underlying architecture matters.
Large language models are remarkably good at reasoning across complex inputs. Deterministic systems are exceptionally good at consistently applying defined rules and producing precise outcomes.
The opportunity is not to choose between them, but to bring them together.
At Avalara, we have spent more than two decades building tax and compliance infrastructure designed to apply complex rules across enormous volumes of transactions and data. AI lets us make that intelligence more accessible and actionable without giving up the trust and precision businesses require. The goal is to expand what experts can accomplish, not replace their judgment.
Reimagining the Work
This is the part I find most interesting.
Every major technology shift starts by getting applied to how people already work – an existing process made faster or cheaper.
The bigger opportunity emerges when the process itself gets reimagined around what the technology can uniquely do.
I believe we are entering that stage with AI.
For tax and finance leaders, that means asking different questions. Which work should stay human? Where should agents recommend an action versus take one? What controls need to exist? And how do we make every action understandable and auditable? These are organizational questions as much as technical ones. The companies that get this right will redesign work around the respective strengths of people, intelligent agents, trusted data, and domain-specific systems, rather than measuring progress by how many AI tools they deploy.
That is particularly important in tax and compliance, where judgment, expertise, and accountability will continue to matter. AI’s job is to give experts more capacity to apply those strengths where they create the most value.
A New Era for Tax and Compliance
We are still early. There is a lot to learn about how agents should operate, how organizations should govern them, and how humans and intelligent systems will work together.
But the direction is becoming clearer.
Compliance systems will become more intelligent. They will understand more context, identify issues earlier, help orchestrate complex workflows, and with the right controls, increasingly help businesses take action.
For businesses, the result should be more consequential than greater efficiency. It should mean greater confidence. Confidence that as your business changes, your compliance systems can change with it. Confidence that complexity does not have to become a barrier to growth. And confidence that AI can be put to work in ways that are powerful, responsible, and trusted.
These are some of the questions we will explore at Avalara CRUSH North America in Fort Lauderdale, September 22–24, and Avalara CRUSH Europe in London on October 8. I look forward to hearing how customers, partners, tax professionals, technologists, and business leaders are thinking about this transition and sharing more about where we believe tax and compliance are headed.
The most important change may be how much complexity AI can absorb.
If we get this right, businesses will move faster and grow while more of that complexity gets handled intelligently in the background. That is the promise of the agentic tax and compliance era.