IBM i’s AI Advantage: The Ecosystem Stakes a Claim
As a platform for AI, the OS has a leg up due to its comprehensive data sets, integrated architecture and unique user base, IBM i enthusiasts proclaim
The IBM i community has entered the AI race, and they think they’ve picked the right horse.
The topic was inescapable for those attending last month’s POWERUp in New Orleans—both in the education sessions and on the expo floor. In fact, AI has embedded itself in the IBM i ecosystem so thoroughly that IBM i shops will be using it whether they actively choose to or not, Steve Will, IBM i CTO, said as he reflected on the conference.
“In the expo area, every single vendor tool or solution had AI built into their stuff, every single one of them,” Will told TechChannel. “So you are going to get AI value as an IBM i client if you use these tools, period.”
IBM i proponents argue that the platform is particularly well equipped to help organizations capitalize on the current AI moment, due not only to inherent characteristics of the system, but to the types of shops and technologists that use it.
Integrated Architecture Supports Complete AI Data Set
The IBM i platform has tended to change more slowly than others, providing a continuity that may now be paying off as organizations realize that the success of their AI projects is influenced by the depth of their data, said Brian May, VP of product management at Profound Logic. May experienced IBM i’s AI prowess first-hand while using the Profound Logic’s own internal agentic coding assistant, which would later be released to customers.
“The key to having good analytics and good AI is having enough data for it to actually be able to infer,” May told TechChannel. “And if there’s anything IBM i shops have, it’s data, many times 40 years’ worth of data, just sitting there waiting to be used.”
Will noted that due to the integrated nature of IBM i, those shops always collect their data. Meanwhile those on distributed platforms may be dealing with data scattered across silos, he added, or may not even collect it at all—perhaps because their OS isn’t inherently responsible for it and nobody has set up the tools to capture it.
Some of that data is about the system itself, such as the performance data that IBM i automatically stores as part of the platform’s emphasis on performance management. All that historical data can then be leveraged in AI-driven performance management tools, Will points out.
Put simply, the architecture of IBM i requires less wrestling with data and system components, he explains. “This architecture that we have covers so much of what AI needs, not only for the data it’s going to use, but for the protection, management, etc., that are going to be behind AI,” Will says.
During the opening session of POWERUp, IBM Power GM Hillery Hunter made IBM i’s advantages as an AI platform the central theme of her remarks. “There are very few other systems that are end-to-end integrated,” she said, contrasting that with the challenges faced by larger enterprises on distributed platforms.
Those organizations have to build their AI foundation from scratch, Hunter noted, and they have a series of questions to answer: Where does the data lake go? What data warehouse should we use? What’s our governance solution? What do we use as the middleware that stitches it all together?”
Compare that to IBM i, Hunter said, “where all the pieces are effectively predefined, prepackaged and ready to go.”
SMBs Poised for AI Adoption
As large enterprises continue to struggle to get ROI out of AI projects, IBM i shops, by comparison, may be ready to pounce.
“There’s an incredibly unique opportunity to move at the fastest pace in the industry,” Hunter said, “and I truly believe that with IBM i, because there are very few other systems that are end-to-end integrated.”
The speed at which IBM i shops can deploy AI isn’t all due to the platform. It also has to do with the kind of businesses that tend to run IBM i. The platform is deployed by organizations large and small, but is best known for serving as the all-inclusive computing backbone of small- to medium-sized businesses (SMBs).
Larger organizations have massive infrastructures across which to scale AI. And while the technical challenges, like data fragmentation, are numerous, successful implementation may also be a matter of bureaucracy.
“If you are a tire shop owner or you are a library or you are a sheriff’s office, you’re going to have one or two specific things that you want to do and they’ll be fairly quick to implement, and then you’ll get value fairly quickly,” Will said. “If you’re a larger company, you probably have rules about who you can interact with and how these different departments are going to have to work with each other and so on. It can take them longer.”
The typical IBM i shop may also have more to gain from AI as a matter of proportion. For larger organizations, it’s a common pattern: Early results look promising, but when the organization attempts to put its pilot AI system into production, progress slows or comes to a complete stop. In that scenario, AI isn’t a force multiplier, but instead a complexity multiplier.
The typical IBM i shop doesn’t face the same challenges of scale. They might not be able to quickly staff up like larger enterprises, but they might be nimble enough to beat them to the punch in deploying AI solutions.
“I think our SMB clients have the most potential value to gain because most of them are running with very small staffs, and technologies like this will allow them to scale without scaling headcount,” May said.
Added Will, “I think that if they take the step, they will get there faster than the giants and see results more quickly.”
AI-Assisted Development, Specialized for IBM i
One option for shops looking to inject AI into their enterprise is IBM Bob, the new AI-powered “development partner.” The platform-neutral Bob is designed to assist along the entire software development lifecycle, but will gain more utility for IBM i users when the IBM i “premium package” is released, expected some time in the second quarter.
The special IBM i package will allow Bob to reach directly into the IBM i file system, meaning developers won’t have to move source code in and out of their native IBM i environment. That will alleviate one friction point for shops used to the IBM i’s unique source member structure, Will explained.
The IBM i package will also include prompting assistance specially tuned for IBM i. This, Will said, will help organizations save on token costs by enhancing efficiency.
It will also reduce IBM i shops’ hesitance to tinker with their decades-old code, Hunter said.
The goal with Bob, she explained, “is to remove that fear factor in the IBM i community—that fear factor of the code that hasn’t been touched, the fear factor of the code that’s in RPG II or III, the fear factor of the data access methods not being fully modern.”
ISVs are taking advantage of Bob, too, integrating their own AI tools with the product. For instance, Profound Logic’s AI coding assistant, CoderFlow, is an agentic layer that can sit on top of IBM Bob or the large language model of the shop’s choosing, such as Claude, Gemini or ChatGPT.
Will said that during the development of Bob, he was sure solutions providers would find ways to use Bob as a foundation for their own products. “Particularly the folks who do code development, software development lifecycle stuff, modernization—they’ve been watching Bob and seeing what Bob is going to do as a base product, and they have found ways extend it, to include their specific flavor of doing whatever they do in the software development lifecycle,” Will said.
AI-Ready Technologists
If IBM i shops have an AI advantage, it’s not just due to their platform; it’s also the people that run it.
IBM i developers tend to stay with their platform longer than others, May said. More time in a particular IT environment means more knowledge of that environment, context that comes in handy when crafting AI prompts.
But it’s not just system knowledge that matters. Business knowledge is the other side of the coin, and IBM i developers are in a unique position to soak that up, too.
Developers on most platforms tend to be specialized, but on IBM i, this is less often the case, particularly in smaller shops where developers may fill many roles, May said. Those technologists are not just writing code, but have institutional memory of the business operation.
“In the IBM i space, you’re not an RPG developer or a COBOL developer. You’re really a business analyst,” May says. “You have that tribal knowledge about how the business actually works. And that’s extremely valuable when you’re talking about building AI solutions—being able to give a little extra context, a little extra understanding that you already have in your head, and getting that into your prompt so that the results you get out of AI just are that much better.”
IBM i developers may also simply have more time to focus on the business side of things. IBM i programs tend to be less complicated than others, May says, leading to less babysitting and fewer “fires” to put out.
“Generally, you write an RPG application, a COBOL application, it just works,” he explained. “That opens up the opportunity to get a little more involved with the business side as well.”
At POWERUp, Hunter told a ballroom full of those IBM i users that they have an AI-ready platform already at their fingertips. “The question,” she said, “is really what you do with it.”