Building Search for AI, Not Humans | Saahil Jain, CTO of You.com
Search has always been built for people—but what happens when the user is an AI agent? In this episode of Ship Happens, host Per Krogslund sits down with Saahil Jain, CTO of You.com, to explore why the next generation of search needs to be designed for machines, not humans. As AI agents become increasingly capable of researching, reasoning, and completing complex tasks, traditional search engines aren't equipped to meet their needs. Saahil shares how You.com evolved from a consumer search engine into enterprise AI infrastructure powering more than one billion monthly queries. Together, they discuss why search built around human behavior breaks down for AI agents, how enterprises are moving beyond AI experimentation into operational AI, and what it takes to build reliable systems that balance speed, accuracy, and cost. From evaluation frameworks and multi-model architectures to continuous monitoring and the future of enterprise AI, this conversation offers an inside look at the infrastructure powering the next generation of intelligent software. What You'll Learn: • Why traditional search engines weren't designed for AI agents • How You.com is rethinking search for an agent-first future • What separates AI adoption from true AI operations • How enterprises balance cost, latency, and model accuracy • Why evaluation frameworks are becoming critical AI infrastructure • The benefits and challenges of multi-model AI architectures • Why continuous monitoring matters as AI models evolve • Where the next competitive advantages in AI will come from Chapters: 00:00 The Agentic Search Boom 00:43 Meet You.com’s CTO 01:39 Building Search for AI Agents 02:45 From Consumer Search to Enterprise AI 03:59 Why AI Agents Can’t Just Use Google 06:55 Rethinking Search for Enterprise AI 07:39 Giving Agents Control Over Search 09:05 Where Companies Are on the AI Adoption Curve 11:42 AI-Native vs. Traditional Companies 13:11 The Real Metrics Behind AI Operations 14:46 Why AI Evals Need to Go Beyond Benchmarks 17:13 Building for a Multi-Model World 19:42 Shipping AI: Gates, Drift, and Monitoring 21:17 Rethinking Org Design for the AI Era 22:53 What Actually Wins in AI Products 25:33 Avoiding AI’s Biggest Engineering Trap 27:58 Where the Next AI Moats Will Come From 30:55 Closing Thoughts If you enjoyed this episode, subscribe to Ship Happens for conversations with the engineers, founders, and technology leaders building the future of software—from AI and platform engineering to cloud infrastructure, security, and developer experience.