Kyndryl Launches AI Framework to Revitalize Legacy IBM Mainframes

Kyndryl is betting that AI agents can do what years of modernization efforts haven’t: Move the most entrenched mainframes into a faster, more automated era without disrupting the critical workloads they carry. The company is launching an AI-driven framework built around intelligent agents and its Kyndryl Bridge platform, initially supporting IBM’s z/OS mainframes.

“Mainframe teams face pressure to move faster and operate more efficiently, but skills gaps and complex workflows make that hard,” Hassan Zamat, Kyndryl’s global practice leader for core enterprise and zCloud, explains to The AI Economy in an email. “Our agentic AI services tackle these challenges head-on by automating repetitive tasks, speeding up decision-making, and giving teams access to decades of expertise through AI-powered tools. The result? Faster issue resolution, better reliability, and the ability to innovate without adding risk or cost.”

Modernizing mainframes is rarely simple. Replacing these systems can be prohibitively expensive, and IT teams must also navigate complex legacy code, difficult integrations, massive on-premise data, security demands, and the shrinking pool of staff trained to operate these systems.

Can AI address these hurdles? Kyndryl’s framework includes several core components:

• Kyndryl AI Assistant for Z: This is an AI-powered knowledge base bot trained on Kyndryl’s mainframe experience. It’s integrated with Kyndryl Bridge and provides support around platform management. This bot is built explicitly for this IBM use case.
• Kyndryl Bridge: This open-source platform enables intelligent automation, predictive analytics, and insights across mainframe hardware and software.
• IBM watsonx Assistant for Z: Launched in October, this agent powers data-driven intelligence and workflow automation for IBM’s Z mainframe.

Additional tools handle data ingestion, orchestration, and execution, enabling multiple agents to work seamlessly across mainframe environments.

“We focused on IBM Z because that’s where our customers run their most critical workloads and where automation needs to be rock-solid,” Zamat says. “Kyndryl manages the world’s largest base of IBM Z systems, and our close relationship with IBM lets us bring AI-powered insights right to the data on the platform.” There’s also a sentimental reason for the move—Kyndryl’s existence as a standalone company came after being spun off from IBM in 2021.

Kyndryl describes this integration as a “significant milestone” for the company. “By weaving agentic AI directly into how we deliver services, we’re making those systems more innovative and more responsive,” Zamat states. “It means faster decisions, fewer manual steps, and more reliable operations. Most importantly, it shows that AI isn’t just an add-on; it’s becoming a core part of how organizations manage and evolve their most critical systems.”

He asserts that mainframes are uniquely positioned for reinvention through AI, especially since they’re the backbone of data-intensive industries like banking, retail, and healthcare. “Bringing AI directly to the platform improves performance, security, and customer experience. By placing AI-powered analysis close to mainframe data, organizations reduce delays and unlock new use cases, like real-time fraud detection or smarter claims processing.”

Traditionally, organizations would have to buy new mainframes to access these capabilities—existing hardware doesn’t have the required technology. However, frameworks like Kyndryl’s bring modern features to existing systems, revitalizing legacy infrastructure so they’ll be relevant for a little bit longer and save organizations money.

Zamat offers this reminder: Most critical data still lives on mainframes like IBM Z. The addition of AI on this hardware promises to ensure the systems handling revenue and compliance are just as intelligent and automated as the rest of an organization’s workflow.

“Look for solutions that integrate with what you already have instead of requiring major overhauls,” he advises IT leaders. “Make governance a priority: AI should be explainable, secure, and aligned with your processes. Choose partners who bring both deep operational expertise and AI capabilities. Success isn’t just about algorithms—it’s about applying them in ways that work for your business.”

Featured Image: Credit: Kyndryl