Expedia Group at Skift Data + AI Summit 2026

Skift Data + AI Summit 2026

THE ARGUMENT

Ranganathan argued that at Expedia driving customer engagement and great outcomes is more important than which tech model wins. An important issue is when to walk away from a product introduction that didn’t work. An early chatbot that wasn’t grounded in Expedia’s own inventory and rates didn’t get traction because customers didn’t find it relevant. So Expedia abandoned the product. Expedia recently produced a property Q&A feature with three times the engagement of the tool it replaced. On build versus buy, her answer was structural: Anything touching customer data and personalization is a build. There’s no need to build or buy service tools such as Salesforce or AWS because they are doing a good job. Expedia will consider buying when a company is doing something great for a particular use case. Ranganathan said the barrier to AI adoption is organizational, not technical, and that the fix is removing swim lanes so that product managers and others can take work from idea to near-production code — and no one is territorial about the result.

THE SO WHAT

If you’re setting your AI roadmap around build versus partner or buy, Ranganathan’s framework suggests a major issue is whether the capability touches customer data and trust. If it does, outsourcing it is a structural risk. On the LLM booking threat, she argued the right posture isn’t defensive: Partner with the platforms your customers are already using and make sure your inventory shows up there, because customers will dictate the channel.