IT modernization has been a perennial goal, but a new survey from Kyndryl suggests AI is not yet the silver bullet. Only 10% of senior IT decision-makers say they are using agentic AI as a modernization tool, and many existing efforts are failing to deliver. The survey of 2,000 IT leaders found that only about half report better IT operations after modernization, and less than half see innovation gains; 18% saw limited or unclear value, and nearly half are behind schedule or over budget.
Andy Thurai, a former IBM strategist, says the economics don't hold up. He argues that AI is the latest shiny distraction from untangling technical debt, and that AI-driven infrastructure sprawl can turn monthly bills into a black box. Three in ten leaders already worry that AI-driven modernization will end in wildly unpredictable compute costs.
Still, agentic AI shows early promise in specific areas: mapping hidden dependencies between legacy and new systems, generating code, and creating documentation. Organizations not using it are much more likely to be behind schedule than early adopters. But Thurai warns that today's AI-driven architectures will eventually become the legacy environments somebody else has to modernize, making the quest a chronic condition rather than a one-time fix.