In its 2026 Mainframe Survey launched this week, BMC revealed how firms are utilizing AI with mainframes. The info signifies a transparent change in how companies use this know-how, with the main target shifting from testing AI to utilizing it in day by day work.
The survey gathered solutions from greater than 1,300 professionals and decision-makers around the globe. The outcomes present that mainframe know-how stays central to enterprise. About 94% of respondents stated they’ve long-term confidence within the platform and stated they plan to proceed investing in it. For 45% of these surveyed, implementing AI is a prime precedence.
The Transfer to Trusted AI
For years, AI on the mainframe was usually experimental. That section is ending. Firms now view AI as an operational instrument. Nonetheless, the info exhibits that this adoption is pragmatic. Organizations are cautious. They need AI to supply perception and recommendation, however they aren’t prepared to provide it full management.
This cautious realism is obvious within the survey numbers. For instance, 40% of respondents are keen to have AI recommend actions for code administration, but solely 23% are snug letting AI full these actions by itself.
An analogous pattern seems in database reorganization. About 43% of respondents need AI to advocate actions, a rise from 37% the earlier yr. Regardless of this curiosity in suggestions, solely 21% of respondents are snug letting AI end the duty with out human assist.
Targets and Hurdles
Firms are prioritizing AI initiatives that provide clear advantages – enhancing productiveness and simplifying their operations. In addition they need to use AI to protect information and velocity up the modernization of their techniques. Widespread duties for AI embody efficiency tuning, drawback detection, managing IMS queues, and producing documentation.
Regardless of these objectives, groups face obstacles. The survey highlights a number of considerations that decelerate implementation:
- Implementation prices: 41% of respondents cite excessive prices.
- Safety and privateness: 39% establish these as main dangers.
- Information integration: 37% be aware that shifting and utilizing knowledge is troublesome.
- Regulatory and compliance guidelines: 22% discover these necessities a barrier.
Future Investments and Safety
Firms are looking forward to agentic administration. This includes utilizing AI brokers to handle elements of the mainframe. The survey exhibits that 36% of organizations plan to construct their very own brokers. One other 32% plan to purchase brokers from third-party distributors.
Safety administration can be a particular space of focus. New guidelines require lowering the usual digital TLS certificates life cycle to 88% by 2029. This variation will create extra work for IT groups, and in the event that they don’t have a transparent plan for that, these groups might face service disruptions.
At present, certificates administration is cut up throughout the trade. About 43% of retailers use automated options they constructed in-house. One other 31% use industrial, product-based automated options. Nonetheless, 25% nonetheless depend on guide effort. This hole represents a threat as new compliance necessities take impact.
The Human Function
John McKenny is the senior vp and common supervisor of Clever Z Optimization and Transformation at BMC. He stated the trade is shifting previous the stage of easy experimentation. He notes that firms have stopped asking, “How can we use AI?” and at the moment are asking, “The place can we belief it?”
McKenny believes that the trail to AI autonomy on the mainframe relies on belief. He stated that companies must know the place AI will be ruled and the place it delivers actual worth. He emphasised that the choice for people to remain within the loop will stay mandatory. People will proceed to supervise and implement suggestions from AI instruments.
