Inside Avalara’s Agentic AI Tax Compliance Platform
Agentic AI tax compliance is quickly becoming a real category. Hugo Sarrazin, CEO of Avalara, joins Alan Shimel on Techstrong TV for a second conversation. Furthermore, he unpacks how Avalara wraps deterministic tax calculation with agentic AI without breaking accuracy.
About Hugo Sarrazin
Hugo recently stepped into the CEO role at Avalara after a technology and product career that included Ultimate Software. In addition, Avalara is a cloud first, multi cloud platform that runs on Google, AWS, Azure and Oracle. Consequently, the company powers tax calculation on more than two trillion dollars of transactions.
Why agentic AI tax compliance matters
Hugo explains that the deterministic calc engine is unchanged and delivers 99.9999 percent accuracy at low latency. Meanwhile, agentic AI tax compliance now wraps that core to monitor Nexus thresholds and prompt exemption renewals. Therefore, buyers get automation without giving up audit ready determinism.
Avalara uses 22 years of proprietary transaction data to train specialized agents. Furthermore, the platform closes a full feedback loop across calculation, filing and audit outcomes. As a result, agentic AI tax compliance can anticipate problems before a return is submitted.
MCP, A2A and living in other surfaces
Hugo notes that Avalara has run on MCP for a while and uses A2A for agent to agent workflows. In addition, the same engine now lives in Slack, email inboxes and ERP agents. Consequently, agentic AI tax compliance meets teams where the work already happens.
He also flags a coming world of AI to AI commerce with no human in the loop. Meanwhile, Avalara is preparing a tax framework for agent driven transactions. Therefore, enterprise leaders should plan for agent to agent settlement, not just human checkout.
Crush 2026 and what is next
Hugo previews Crush 2026 in Fort Lauderdale on September 22 to 24 with his first Avalara keynote. Furthermore, tickets are still available for teams that want to see the roadmap in person.
Explore more artificial intelligence coverage and the latest Techstrong TV interviews.
For more information please visit avalara.com
