
Agentifying the Mainframe: Where Agentic AI Goes Next
transcript
show notes
Everyone wants AI on the mainframe. Very few want to hand it the keys. I sat down with Tarun Chopra, SVP & GM at BMC Software, in New York on The Ravit Show, the next week BMC announced BMC AMI Agentics, a new product line as part of their vision for "Agentifying the Mainframe."
What stayed with me:
Trust comes before autonomy. BMC's 2026 Mainframe Survey found willingness to let AI complete actions in problem diagnosis dropped from 32% to 23% in a year. Preference for AI recommendations rose from 36% to 41%. Tarun sees that as teams being careful about where trust belongs. I agree with him.
Generic AI pointed at mainframe data won't get you far. The agent needs decades of mainframe expertise, your team's own knowledge, and what's happening in the system right now.
The real bottleneck is people. A lot of mainframe knowledge sits with a handful of experts. Agents that investigate, explain and recommend let the rest of the team use that knowledge.
And customers are clear: the mainframe is staying. They want it to be part of their enterprise AI plans.
Full conversation below. Where do you stand: recommendations first, or ready to let agents act?
#mainframe #agenticai #ai #bmc #enterpriseai #aiagents #technology #theravitshow