AI Won't Fix Legacy IT Economics, Former IBM Strategist Says as Kyndryl Survey Shows Slow Agentic AI Adoption

IT modernization has been on CIO agendas for decades. Efforts have included web services, microservices, cloud, virtualization, and x86-based Windows and Linux servers, all aimed at creating more advanced and flexible systems. But Kyndryl researchers found that modernization is often just another expensive, time-consuming project. Only about half of those modernizing report better IT operations, and less than half are seeing innovation gains. Nearly one in five, 18%, say their firms saw limited or unclear value from modernization efforts. Almost half of respondents said they are behind schedule and suffering cost overruns.

Agentic AI could shape the next wave of modernization. Instead of replacing mainframes and other legacy systems, the focus is on expanding tangible business capabilities. Kyndryl reported that cost reduction and legacy escape are no longer the leading motivations to modernize. As AI increases demand on nearly every element of IT, and as sovereignty considerations become more acute, leaders want their complex technology estates, a mix of old and new decisions, to work better.

Agents can operate autonomously while providing more visibility on dependencies between systems, both legacy and new. Fewer than one in 10 organizations are fully confident they know those dependencies today, a significant technical blind spot, the Kyndryl report stated. Kyndryl authors added that agentic AI is showing early promise to address longstanding challenges in modernization and speed progress. Agentic AI is being used to map hidden dependencies, generate code, and create documentation, all traditionally time-consuming, error-prone processes. Organizations not using agentic AI for modernization are much more likely to be behind schedule than early adopters.

Modernization challenges remain despite AI's promises. Andy Thurai, founder of The Field CTO and former chief strategist with IBM, said IT modernization increasingly looks like a chronic condition. According to the report, four in ten enterprises are still running mainframes. "We keep hearing modernization is the number-one mandate, yet the finish line is always 5 to 10 years out," Thurai said. Another issue, he said, is that organizations are addicted to buying software. The legacy systems underneath it almost never get decommissioned. AI is now the latest shiny distraction from the grind of untangling that technical debt.

CIOs must contend with vendors pitching AI to IT leaders as the cure-all, Thurai continued. "The economics don't hold up," he said. When AI drives infrastructure sprawl, the monthly bill becomes a black box. The report found that three in 10 leaders already worry AI-driven modernization will end in wildly unpredictable compute costs.

AI-based modernization may have its moment, especially for organizations that went past the bolt-on chatbot, Thurai said. They are using agentic frameworks as an orchestration layer to map out hidden technical dependencies, and then generating the missing documentation as they go. Agentic AI is speeding up initiatives that used to eat months of error-prone manual labor.

Sooner or later, today's AI-driven architectures are going to be the legacy environments somebody else has to modernize, Thurai said.