What happened

Rocket, based in Waltham, Massachusetts, announced the deal in a press release distributed by GlobeNewswire. It describes Geniez as a generative AI company whose agents monitor live systems, diagnose incidents and carry out workflows from instructions written in everyday language. Geniez's technical leadership and engineering team are joining Rocket.

Rocket did not disclose the price. The Israeli business site Calcalist reported estimates of up to $100 million for the six-person company; neither company has confirmed a figure. Calcalist also reported that Geniez raised a $6 million seed round in September 2025, and that its founders, Gil Peleg and Dan Shprung, had earlier founded Model9, which BMC Software bought in 2023.

Rocket says the deal adds AI and mainframe talent, speeds up its agentic AI roadmap and helps close the mainframe skills gap. It also says it wants to reduce operational complexity through governed, AI-assisted operations.

Why it matters

Rocket is a large vendor. Its release says it has more than 13,500 customers, 850 partners and more than 3,000 employees, and that it is a portfolio company of Bain Capital Private Equity. When a vendor of that size buys an agent specialist, more mainframe sites are likely to meet AI agents inside the tools they already use.

Some questions are still open. An analysis by BERI notes that Rocket has not said whether Geniez will stay available as a separate product, what it will cost, or when the new EVA capabilities will be generally available. Existing Geniez customers will want clear answers on support and contract continuity.

There is also a wider point for operations teams. Modernization rarely happens in one move. Workloads leave the mainframe over months or years, and the batch that stays behind still has deadlines. During a migration, it is easy for attention and budget to follow the new platform, while the old estate keeps running the payroll, the statements and the settlement files.

Questions to ask before an agent touches production

Whichever vendor you choose, these questions are worth asking before any agent, human-guided or not, acts on a production LPAR:

  • Who decides? What can the agent do on its own, and what needs a person to approve it?
  • What is recorded? Is every action logged, along with what the agent saw when it acted?
  • Where does the data go? Does any operational data leave your network, and to which service?
  • Does it know the deadlines? Can it tell which failed job threatens which batch SLA, so the team fixes the one that matters first?
  • What if the product changes hands? If a tool is folded into a larger suite, what happens to your support, pricing and roadmap?

A job that abends at 2am still has to be fixed or restarted before the business day starts. Any new tool should make that night easier to run, not harder to explain the next morning.

In the words of Milan Shetti, president and CEO, Rocket Software: "This is about continuing to transform what is possible on the mainframe."

How Meridian helps

Meridian is Hexabin's mainframe operations and batch monitoring product for managed service providers and enterprises that run z/OS. It watches every LPAR and says in plain words when a batch will miss its SLA, hours before it does. It acts through cards that an operator confirms, and every action is recorded with its evidence. Meridian runs inside the customer's own network, and the customer's data never leaves it. Teams can try it with a free 90-day proof of concept on one LPAR.

The bottom line

AI agents are arriving in mainframe tools, and Rocket's purchase of Geniez will bring them to more of its customers. The test for any of them is the one operators already apply to each other: did the batch finish on time, and can you show exactly what was done to get it there.

Sources

  • globenewswire.com
  • finance.yahoo.com
  • calcalistech.com
  • beri.net