Rocket Automates Manual IBM i Tasks With AI

September 14, 2026 Alex Woodie

Rocket Software has a multi-pronged AI strategy that spans multiple products and platforms. When it comes to the IBM i platform, this AI push has resulted in Rocket adding AI features that will automate previously manual tasks across its application modernization, change management, and high availability products.

AI continues to be the elephant in the IT room, but unlike those quiet elephants in that expression, which no one admits is there, this is the elephant that nobody can stop talking about and for good reason: The 2026 edition of AI appears to be the real deal.

Take a KPMG survey from March for example. The Global AI Pulse Q2 Survey found that 95 percent of C-suite leaders are executing an AI initiative, which is as close to ubiquitous as one can get. What’s more, 76 percent of the 2,100 execs who took the KPMG survey said they “see real business value from AI,” while 71 percent say they are progressing toward “a fully integrated AI-human workforce.” Both of those measures were up more than 10 percentage points from the first quarter.

As AI technology improves and business leaders get more comfortable with it, they will use it in more ways. Rocket Software is following this pattern as it seeks to layer AI into its commercial software products. The Waltham, Massachusetts software and services firm is taking a cross-platform, cross-product approach to AI that manifests with the “infusion” of “actionable, explainable, and governed” AI into its products for application development, database management, and system operations.

Rocket has recognized the lack of technical skills as one of the biggest issues facing IBM i shops (which was the big takeaway from Fortra’s 2026 IBM i Marketplace Survey), and so it’s looking to AI to give its customers a boost in this area. To that end, Rocket recently announced that it’s using AI to help customers automate tasks that were previously done manually across its DevOps, Continuous Modernization, and iCluster products.

Puneet Kohli, the president of Rocket’s business application modernization unit, recently shared details of the AI infusion in the three products with IT Jungle. For Rocket DevOps, one of ways it’s using AI is generating code test cases.

“Our initial view is that it can increase code coverage beyond 90 percent,” Kohli said. “We can go module by module and recommend unit tests that one can generate. Before you would have to know this. Now it’s AI-powered.”

Rocket is also using AI to analyze the quality of code as part of a static code analysis test, as well as for converting from fixed-format RPG to free-format RPG. These capabilities are part of a unified group that will reduce the need to go to other tools.

“We combine our test generation with the code conversion capability, so now everything is in one place,” Kohli said. “You don’t have to go from one tool to another tool. You’re in Rocket DevOps. It’s one-stop shop for doing your test conversion as well as your code conversion.”

The company also added AI capabilities to Continuous Modernization, which is a combination of the former Rocket API product as well as LegaSuite, which Rocket obtained with its acquisition of Seagull Software 20 years ago. Rocket has added a handful of new AI capabilities, including workflow analysis.

According to Kohli, AI will now analyze the workflow, including things such as the specific 5250 green screens that users are working in and the various database calls that an application is making. “We can now augment all of that and create a much more visual way for you to describe all of those things,” he said.

Rocket has also added a chat interface to the product that allows users to ask questions about the workflow, such as how many times a user is hitting a particular screen or how many database calls are being made. The AI model used by Rocket essentially interprets the business logic of the RPG or other source code, and uses retrieval augmented generation (RAG) methods to enable users to ask questions of it, as opposed to poring through documentation (if it even exists).

The AI capability, which plugs into Rocket’s Eclipse-based IDE for Continuous Modernization, makes the developer interaction with the tool much more conversational, Kohli said. It also eliminates the need for a user to be a 20-year expert

“You don’t need a twenty-year skilled IBM i person now because you can chat with the application,” he said.. “It accelerates the process, makes it a lot faster. Before you would have to know which screen, what widget you need to move. Now it can make recommendations to you.”

Similarly, the new AI-based tuning feature that Rocket has added to iCluster reduces the need for skilled IBM i personnel to monitor the high availability environment and make configuration changes to improve performance.

“That’s a clear example of an admin sitting there having to figure out journal entries and activity peak and traffic patterns to say, OK, how much do I tune my HA/DR process?” Kohli said. “Now we have an AI tuner. You plug that into an agentic flow to get that information, and then based on that, we just make a recommendation that these are the things that you can tune in the product.”

For instance, the AI Tuner could make recommendations about how to organize replication groups. There could be temporary objects on the production IBM i server that could be excluded from replication, thereby speeding up replication, Kohli said. Object dependencies are another area where human administrators must make decisions that impact not only performance, but whether a disaster recovery will be successful.

AI may eventually be making these decisions on its own, but today it is conducting the analysis and providing recommendation, Kohli said.

“Our vision is we’ll get there, but today we are just at the recommendation phase,” he said. “We still want a human in the loop, because these are critical changes. This is where how much you trust the AI comes in, right? Even though we have checks and balances in place, we have the right security and governance in AI. But are our admins comfortable saying, yeah, let the AI go tune my environment? We’ll have to wait to see what kind of feedback we get.”

Rocket also announced new capabilities for its IBM Z, MultiValue, and Uniface products, which run on mainframe and open systems. In some ways, the AI development is further along in the mainframe world at Rocket. For instance, the company has developed an Enterprise Virtual Assistant dubbed EVA for the mainframe. The AI roadmap calls for EVA or EVA-like capabilities to be brought to IBM i eventually, Kohli said.

“EVA is a holistic strategy,” he said. “We don’t plug into that fully yet [with IBM i products], but . . . our roadmap down the road is that yes, we’ll plug all that in.”

This Issue Sponsored By

Table of Contents

Content archive

Recent Posts

Subscribe

Pages

Search