The Agent Development Life Cycle — Zack Reneau-Wedeen, Sierra

Compared to traditional software, LLMs are creative, flexible, unpredictable, expensive, and slow. A new kind of software demands a new approach to development. At Sierra, we’ve built industrial grade AI agents for consumer brands like Sonos, ADT, and SiriusXM, serving millions of consumers. Every day, their agents are reasoning through complex workflows and helping customers set up speakers, order equipment, and refresh in-car radios. As we’ve scaled with our early customers, we’ve developed a unique methodology to enable Sierra agents to be reliable, testable, and incredibly capable. We've turned scale into an asset that helps align goals, enforce guardrails, and enable continuous self-improvement, even and especially as new models, price points, modalities, and paradigms like reasoning have added complexity and made new experiences possible. Recorded live at the Agent Engineering Session Day from the AI Engineer Summit 2025 in New York. Learn more at https://ai.engineer and purchase tickets to our next event, the AI Engineer World's Fair, in SF June 3 - 5 here: https://ti.to/software-3/ai-engineer-worlds-fair-2025 About Zack I've been a product manager, executive, and software engineer for over ten years at Google and other technology companies focused primarily on consumer products. I started working with neural networks in 2015, when we were building the first version of Google Lens, and I had no idea that language models would play such a central role in the future. Now I'm building conversational AI agents at Sierra.