Dive Brief:
- IBM unveiled IBM Z Atlas, a product that aims to simplify mainframe modernization using AI skills and agents, in a Tuesday announcement. The product will help teams plan for and automate infrastructure and application changes, IBM said. The company released a public preview waitlist for IBM Z Atlas Tuesday and plans to make the product generally available in Q1 2027.
- IBM Z Atlas is intended to accelerate customer efforts to modernize IBM Z mainframes, using AI to map the tech stack and then evaluate, test and validate changes. As the product maps customer’s environments, it also aims to reduce manual patch management and knowledge sharing efforts, according to IBM.
- “IBM Z Atlas is intended to generate contextual, traceable artifacts representing relevant system state, configurations, dependencies and changes,” IBM said in it’s statement of direction. “These artifacts are intended to provide reusable evidence that can support planning, validation, readiness, drift assessment, governance, compliance and audit activities.”
Dive Insight:
IBM’s plan to double down on AI-enabled modernization marks the company’s latest step to keep customers on the mainframe as AI eases enterprises’ ability to deploy new applications and systems.
The company made IBM Bob — an AI assistant for code modernization — generally available in April. It has been heavily investing in expanding AI functionality within the mainframe, an effort that companies including AWS, Microsoft and Google have also supported by launching AI-enabled modernization tools in a bid to move customers into their environments.
Vendors promise AI is making modernization easier, but leaving mainframes can be a tough choice for CIOs. Six in 10 organizations have had to pause, scale back or abandon modernization efforts in the last year, a survey of 500 IT decision-makers conducted by managed service provider Ensono found.
Moving off the mainframe could prove difficult for IT leaders solely relying on AI to do the heavy lifting. More than 70% of projects started in 2026 will fail as leaders overestimate generative AI’s capabilities, according to Gartner.
IBM is bidding on building AI into its ecosystem to support enterprises looking to scale the technology while remaining on legacy mainframe infrastructure.
“Organizations are recognizing that AI is only as effective as the system around it,” CEO Arvind Krishna told investors during the company’s Q3 2026 earnings call last week.
