How Odessia ships a consumer-grade travel agent with LangSmith Engine

Description

Francis Davidson, founder of Odessia, shares how his team uses LangSmith Engine to build and maintain a visually rich, low-latency AI travel agent, one that can book flights, hotels, and full itineraries in a single conversation. He breaks down the core challenge: dozens of tool calls that have to resolve in seconds, or users drop off. With LangSmith Engine, his team can aggregate thousands of traces, combine them with source code and eval results, and extract actionable insight, right down to opening PRs at the root of issues. Hear his story, and try Engine today at https://www.langchain.com/langsmith/engine How Odessia turns traces into fixes with LangSmith Engine 00:00 Intro: diagnosing agent issues without deep engineering expertise 00:16 Meet Francis Davidson + Odessia's AI travel agent 00:23 The hard engineering problem behind a simple travel experience 00:57 How LangSmith Engine handles trace aggregation at scale 01:21 Merging traces, evals, and source code into one intelligence layer 01:50 Anyone can now open PRs and fix root issues 01:55 Accelerating the agent development lifecycle with LangChain 02:13 Where Odessia is headed Extra resources: LangSmith Engine: https://www.langchain.com/langsmith/engine LangSmith: https://www.langchain.com/langsmith/ LangChain: https://langchain.com/