I spent a day this week down in Fort Lauderdale at the Avalara CRUSH conference. If you aren’t familiar with Avalara, there is a good chance you have encountered its technology while ordering something online. You probably paid attention to what you bought, what shipping cost and when the package would arrive. The sales tax calculation was just another line at checkout.

Making that line correct is a substantial business. So is making sure the seller files the right returns, maintains the right records and meets its obligations wherever it does business. Avalara has built an established leadership position helping companies manage that work.

You might think a company in that position could sit back for a while and collect the fruits of its success. What I heard in Fort Lauderdale was a company doubling down.

Avalara CRUSH felt more like an Orange Crush. The company is embracing agentic AI with the intention of carrying its position in tax compliance into the next era. After interviewing its technology and product leaders, I came away thinking there is a lesson here for a lot of software companies trying to figure out their own AI futures.

Putting an Established Business to Work

Tax compliance is one of those functions that people outside the business tend to underestimate. Apply a percentage to a purchase and send the money where it belongs. How difficult could that be?

Then consider what is being sold, where the buyer is, which exemptions apply and how the seller’s obligations change as its business grows. Add international operations, different tax systems and the business applications that hold pieces of the information. A fairly ordinary commercial transaction can create a surprising amount of work.

Avalara’s scale reflects how much demand there is for help. According to its corporate fact sheet, the company processed more than 54 billion AvaTax API calls and filed more than 6.6 million transaction tax returns in 2025. It also reports more than 1,400 signed partner integrations.

Those API calls are not a count of individual purchases. They are a measure of the activity running through its tax calculation infrastructure. Taken together, the figures describe a company deeply connected to the systems businesses use to sell, account for transactions and meet their tax obligations.

That matters when you start talking about agents doing work. An agent needs access to the right information and a dependable way to take action. Existing integrations, tax expertise and operational infrastructure give Avalara something substantial to build on.

They also create a responsibility. Customers already depend on this software. Introducing AI has to improve the work while preserving the reliability that earned those relationships.

More Than 100 Agents, With a Place to Coordinate Them

At CRUSH, I interviewed Danny Fields, Avalara’s chief technology and customer operations officer, and Jayme Fishman, its chief strategy and product officer. They told me Avalara has already developed more than 100 AI agents.

That is a significant development effort. But the count alone does not tell customers what they can accomplish. The practical question is how those agents work together and fit into the business.

Avalara’s answer is Aviator, the agent hub it announced on Sept 23, 2026. Customers interact with Avi, the orchestrating agent, which coordinates specialized agents. The initial capabilities include mapping products to tax codes, validating exemption certificates and configuring custom rules. The broader group of more than 100 agents should not be confused with what is available in this initial release.

One example from the announcement makes the approach understandable. A customer describes a needed tax rule in plain language. Avi drafts it, tests its effect on historical transactions and flags conflicts. A person reviews and approves the change before it is activated.

Consider the steps in that process. Expressing what the business needs is only the beginning. Someone still has to translate the request into a rule, assess what it would do and decide whether to put it into operation. Coordinating that work is where agents can provide practical value.

The company says the initial Aviator experience is being delivered to existing customers at no additional cost. That gives it an immediate path to introduce these capabilities within established relationships, although it does not mean every future capability will be free.

I see a deliberate business strategy in that decision. Avalara can help customers become comfortable with a different way of working inside a product environment they already know. It does not have to persuade every customer to start over with an unfamiliar provider.

The underlying architecture deserves attention, too. According to Avalara’s product documentation, its deterministic tax engine remains responsible for calculations. Generative AI supports functions such as orchestration, extracting information from documents and synthesizing tax information. Critical actions require human approval.

That division of responsibility makes sense to me. A tax team needs correct calculations, understandable reasoning and records it can defend. Giving people a conversational interface is useful only if the work behind it meets those requirements.

There is still execution to prove. An architecture description does not establish that every result will be correct. But building agents around existing calculation infrastructure, with explicit approval points, is a sensible way to approach a function where mistakes have consequences.

Avalara is also working on how these capabilities reach customers. Its multiyear strategic collaboration with AWS, announced in January, includes solutions built on AWS and access through AWS Marketplace. Infrastructure matters, and so does making procurement fit the way customers already buy technology.

The combination is what stands out: Domain expertise, operational systems, integrations, customer access and an investment in agents that can coordinate work across them. Each element strengthens the usefulness of the others.

The Opportunity Extends Across Borders

Avalara’s international e-invoicing business adds another dimension to that strategy.

An electronic invoice in this context is a structured, machine-readable document exchanged between systems. Depending on the jurisdiction, the process may also involve reporting to a tax authority or obtaining clearance. Requirements vary, creating a substantial coordination problem for businesses operating in multiple countries.

For a company expanding internationally, compliance can become a series of separate projects. Each market brings requirements that have to connect back to the systems running the business.

Avalara’s E-Invoicing and Live Reporting approach uses a single global API and a common data model, with country-specific requirements activated through configuration. The company also announced its placement in the Leaders category of IDC’s September 2026 assessment of worldwide e-invoicing compliance solutions.

The business appeal is clear. Customers want to enter markets and conduct business without rebuilding their compliance connections each time. A provider that can reduce that burden has an opportunity to become more useful as those customers grow.

This is an expansion of Avalara’s existing international business, and it fits the larger ambition I heard at CRUSH. The company wants to handle more of the compliance process across more of the places where customers operate.

For other software companies, there is a useful lesson in how those pieces come together. Years of expertise and customer relationships can be valuable starting points for an AI strategy. Their value depends on what management does with them.

An established provider can use its knowledge of a customer’s work to identify where agents will help. It can connect those agents to systems already doing important jobs. It can make adoption easier through relationships and distribution it has already built. Those are meaningful advantages, provided the company is willing to invest and change how its product works.

Avalara’s customers will ultimately decide whether this effort delivers. They will judge it by accurate results, less manual work, dependable oversight and records they can defend. Agent counts and launch announcements will matter a lot less than that experience.

What impressed me in Fort Lauderdale was the commitment to pursue it while the existing business gives Avalara a strong foundation. The company is putting that foundation to work, developing agents, introducing a hub to coordinate them and expanding its international opportunity.

That was my Orange Crush takeaway. This is what market leaders do when they are on top of their game: They invest in what comes next and press their advantage.