Just Now Possible Building Trainline’s AI Travel Assistant: How a 25-Year-Old Company Went Agentic
Oct 30, 2025
Join David Eason, Principal Product Manager at Trainline; Billie Bradley, Product Manager specializing in AI; and Matt Farrelly, Head of Machine Learning Engineering. They dive into the development of Trainline’s AI Travel Assistant and the unique challenges they faced. Discover how combining AI with deep industry knowledge enhances user experience and the importance of well-structured guardrails. They also discuss innovative evaluation methods and the potential of scalable, real-time support for travelers, even in a traditional company.
AI Snips
Chapters
Transcript
Episode notes
Three-Part Travel Problem Framework
- Trainline split travel problems into knowing status, severity, and solutions to support customers end-to-end.
- Combining domain data with AI enables richer, contextual travel assistance beyond simple alerts.
Choose Agentic Design For Long-Tail Help
- Do build an agentic assistant when you need scalable, flexible, open-ended support instead of rigid decision trees.
- Enable models to synthesize domain data and handle long-tail user questions to build trust and avoid dead-ends.
POC: Orchestrator With Minimal Tools
- The POC used a central orchestrator plus three simple tools: vector DB, terms-and-conditions tool, and a mock refund endpoint.
- The agent reliably chose the correct tool in tests, giving confidence to move to production quickly.