Why Enterprise AI Projects Keep Failing
InfoWorld, Friday, August 28th, 2026
InfoWorld's David Linthicum says AI projects fail because enterprises, not models, aren't ready.
David Linthicum argues enterprise AI failures stem from organizational unreadiness, not weak models.
He identifies six failure patterns: deploying tech before strategy, disconnected pilots that never touch core systems, AI amplifying bad data, agents automating unengineered processes, hidden token and orchestration costs that erase productivity gains, and governance bolted on late enough to kill projects before launch.
His fix: define measurable outcomes, integrate with ERP/CRM systems, fix data first, and build governance in from the start rather than after a successful demo.