Marriott International and Google at Skift Data + AI Summit 2026

SKIFT DATA + AI SUMMIT 2026

THE ARGUMENT

The hard part of enterprise AI isn’t the build. Asked about a study finding 95% of organizations see zero return on generative AI, Romoff disputed the number, and both speakers pushed past it to the real problem: launching a pilot is easy, scaling it to production is not, and what blocks the jump is people and workflow, not technology. Coleman said automating a task as it exists today misses the point, because the worker only does it that way because they have to. The job is to re-engineer what they’re actually trying to achieve. Both said travel’s fragmented, multi-stakeholder data is the true prerequisite, and that much of the available value was sitting there before Gen AI arrived. The session was less about whether AI works and more about why so little of it reaches production.

THE SO WHAT

Your company may be scoping AI as an automation project — speeding up the tasks people already do. But Coleman argued that this locks broken workflows in place instead of rebuilding them, and that the harder constraint isn’t the technology, it’s skilling your team to ask the right questions. Watch the sector order too: cruise and aviation are already converting AI into revenue, while lodging’s consumer-facing layer, including Marriott’s own conversational search, is still arriving.