A short-form read on where procurement conversations are actually going in 2026, based on the themes recurring across this year’s coverage.
Data structure is the AI story, not the model
The conversation has shifted noticeably from "which AI model" to "is our data ready for any AI model." That shift is overdue and it is the right one — a better model pointed at inconsistent material descriptions still produces unreliable output.
Composable is winning over monolithic
Fewer teams want a single rigid suite handling every module to a mediocre standard. More want strong individual modules on one shared data layer, adopted incrementally rather than committed to all at once.
Regionalization is a data problem, not just a strategy one
Nearshoring, friend-shoring, and multi-region AVL management all keep surfacing the same underlying requirement: a shared benchmark and a shared supplier record that travels across geographies, not separate systems per region that cannot see each other.
Corrective action history is moving out of email
More conversations about connecting NCR, CAR, and PIP history directly to the supplier record used during sourcing, instead of leaving it in inboxes disconnected from the next award decision.
None of this is a dramatic reinvention of procurement. It is the same fragmentation problem, showing up in a few new places, with the same underlying fix: structure the data first.