Join us with special guest, Steve Flitcroft, VP of Software at iVendi, for a behind-the-scenes look at how his team assembles and serves fresh customer context to support a production AI chat experience.

In this session, Steve will break down the end-to-end architecture iVendi runs in production using RudderStack, ClickHouse, and Vercel to capture customer context data, assemble it in the warehouse and serve it on demand at inference time. You’ll get real-world patterns, along with a practical perspective on the tradeoffs and lessons learned.
What we’ll cover

• How iVendi captures and assembles customer context from clickstream, operational, and enrichment data
• Why they use ClickHouse as the system of record for customer context
• How Vercel’s AI SDK fits into the serving layer to orchestrate context injection and LLM interactions
• Key design decisions and pitfalls to avoid when assembling and serving customer context for AI

Whether you’re building AI-driven customer experiences, modernizing your analytics stack, or looking to operationalize real-time data, this webinar offers concrete, battle-tested guidance from a team doing it in production.