Rocket Software expands its Rocket EVA platform to deploy AI agents in a controlled manner on mission-critical mainframes.
Rocket Software announces an expansion of Rocket EVA, an AI platform specifically aimed at mission-critical systems on the mainframe. The solution focuses on controlled, auditable AI agents that analyze operational data and automate tasks according to predefined policies. The company aims to address the shortage of specialized mainframe expertise with this move.
Research by Hanover Research commissioned by Rocket Software shows that 81 percent of IT leaders in the financial sector experience a very or extremely significant mainframe skills gap. At the same time, 87 percent believe that AI will help close that gap in the next two years.
Agentic AI on mission-critical mainframes
With Rocket EVA, Rocket Software aims to let AI agents reason directly about mainframe environments in real time. The agents correlate logs, metrics, and operational data and, according to the platform, can automatically suggest or execute actions within strict policy boundaries. This should allow organizations to deploy AI without compromising the stability or security of their core systems.
According to the company, EVA helps teams move faster from problem detection to resolution by bringing together data and context from both mainframe and distributed environments. The focus is on reducing the Mean Time to Resolution through automatic correlation of incidents and accelerating analyses that traditionally required significant manual effort.
PlanGuard
A key newcomer to the platform is Rocket PlanGuard, an additional security layer between the AI agent’s reasoning and the actual system actions. PlanGuard adds a policy decision point and identity checks so that every AI-driven action remains within pre-approved boundaries. All activities are logged to enable audits and human oversight.