The $15M-a-Month AI Bill: Why “Token Maxxing” Is Draining Your Budget

In this episode of How We Agent, Ani sits down with Sri Kandikonda, Group Product Manager at FedEx, who leads the evolution of FedEx’s API ecosystem across 200+ countries. Sri has 15+ years in e-commerce, logistics and fintech, teaches a Maven course on taking AI agents to production, and wrote a book on the API economy.

Sri shows how his team of product managers works alongside AI agents every day. He shares an internal agent that resolves 80% of product questions and explains why “anyone can build an agent, but few can build a reliable one.”

The conversation covers agent guardrails, evals, governance, LLM selection, token budgets, building internal before external, and why AI agents need real-time data to make decisions.

What is How We Agent?

"How We Agent" is Lyzr AI's podcast and video series taking you inside real enterprise AI deployments with sovereign AI, agentic automation, and the production-ready agents actually running in the world's biggest companies. Each episode sits down with industry leaders building (and governing) AI agents at scale