Real-time tax compliance puts agentic AI accuracy to the test
SiliconANGLE Victoria Gayton
Avalara is using AI agents for tax compliance, but only where they can stay penny-perfect. That matters because tax has to be exact across 190+ countries and 87,000 sources.
Based on reporting by SiliconANGLE, Victoria Gayton — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Avalara is putting agentic AI into one of the least forgiving corners of software: tax compliance. The company can’t afford fuzzy answers here. A tax calculation is either right or it isn’t, and Marcus Larner, Avalara’s senior vice president of engineering, kept coming back to that point in his conversation with theCUBE at Google Cloud’s AI Agents in Action Series.
So Avalara is splitting the job in two. Agents handle the front end — taking requests in natural language, deciding what needs to happen next and pulling in the right data — while a deterministic tax engine does the part that has to be exact. In a live demo, the Avi agent validated an address, identified a tax code, calculated tax, mapped liabilities and generated a transaction report. That engine leans on more than 87,000 regulatory sources across more than 190 countries.
This isn’t new territory for Avalara. Larner said the company has been using AI and machine learning for more than a decade, which gives it a pretty practical view of where automation helps and where it can go off the rails. The rule, as he put it, is to apply AI “smartly and intelligently” without giving up performance, trust or accuracy. That’s the real story here: not AI replacing tax logic, but AI being fenced in by it.
Reliability is also a deployment problem. Avalara runs AvaTax in an active-active setup across regions and cloud providers, so customers can move if one location has trouble. That flexibility helped the company place AvaTax alongside Shopify’s workloads in Google Cloud data centers, because Shopify needed the service in the same physical location for performance and resiliency.
The timing matters too. Governments are moving from periodic tax reporting toward real-time systems such as electronic invoicing, while lower compliance thresholds are raising audit risk. Larner said that pushes finance away from pure reporting and toward forecasting, planning and strategy. In other words: tax data is becoming live business data, and that makes correctness a lot less optional.
My take — AI-written commentary, not fact-checked reporting
This is the sensible kind of AI story: narrow the task, keep the math locked down, and don’t pretend a chatbot is a tax attorney. The industry keeps trying to slap agents on everything, but finance still punishes sloppiness faster than marketing ever will. A little less magic, a lot more guardrails — that’s the adult version of agentic AI.
Read more about this at: SiliconANGLE