Hilton and Air Canada at Skift Data + AI Summit 2026

SKIFT DATA + AI SUMMIT 2026

THE ARGUMENT

Both executives said the AI conversation skips the hard part: the foundation underneath. Leidinger described eight years of rebuilding Hilton from a legacy monolith into cloud-native microservices, and Al Osman said Air Canada came off the mainframe last year. The bigger claim was Al Osman’s. Agentic software development is collapsing the cost of building, so his default question is no longer buy-or-build but whether his team can build it in a few weeks. He said enterprise software vendors whose value is arbitraging build complexity face a reckoning. Leidinger said Hilton still buys commodity products and customizes them, treating build-and-buy as complementary. The session was less about whether to modernize than about how far the build-cost collapse moves the line.

THE SO WHAT

If you are funding AI features before the data and event architecture beneath them, this session argued you are building on sand. Both leaders said the companies put in close to a decade on the foundation, but the sequencing differs: Leidinger rebuilt Hilton’s stack first and shipped AI on top, while Al Osman ran AI alongside a cloud and semantic layer he is still completing. The second warning is sharper: if your software budget still defaults to buying from vendors who shield you from build complexity, Al Osman’s claim is that agentic coding is erasing that value fast, and the first question should be whether you can build it in a few weeks.