--- title: "Bob for IBM Z: Bringing Agentic AI to COBOL & the Mainframe with Rich Larin" date: 2026-11-12T09:00:27Z modified: 2026-11-04T18:18:07Z permalink: "https://redmonk.com/videos/rich-larin/" type: video status: publish excerpt: "" wpid: 10893 categories: - Podcast series: - Conversations channel: - RedMonk Video featured_image: "https://redmonk.com/wp-content/uploads/2026/10/Rich-Larin.jpg" timestamp: 2026-11-04T18:18:07Z tags: - Podcast - Conversations - RedMonk Video --- Recorded live at IBM TechXchange 2026, RedMonk’s Kate Holterhoff talks with Rich Larin, Director of Z DevX & Platform Engineering at IBM, about bringing modern, AI-powered tools to mainframe developers. Rich explains why mainframes still power credit card transactions and airline reservations, and how Bob and the Premium Package for Z help teams understand, modernize and extend COBOL applications using the same agentic IDE other developers use. They discuss deployment options, including SaaS, hybrid and air-gapped on-prem, for customers protecting source code. Rich also covers Z code scan, a deterministic security check that Bob can remediate automatically, and a data dictionary that captures experts’ knowledge and shortens onboarding from years to months. He closes with advice to try a proof of concept and a hint at upcoming releases. The conversation also touches on token spend, developer skills, and making mainframe careers attractive to the next generation of developers. This RedMonk video is sponsored by IBM. ## Links - [LinkedIn: Rich Larin](https://www.linkedin.com/in/richlarin/) # Transcript **Rich Larin (00:21)** Hey, thanks so much for having me. Like I said, my name is Rich Larin. I’m with IBM. I’m in the Z software team, and my job is to bring modern tools to mainframe developers. And that’s what I do. **Kate Holterhoff (00:31)** Love that. Okay. I’m so jazzed to have someone with a mainframe background here to talk with us because we talked to a lot of developers at RedMonk.. I think a lot of us we’re very aspirational about hardware, but we don’t always get to to see it and hear about all the cool things that are going on. So I mean I guess just to begin with, what’s doing exciting around IBM Z and what are you seeing around mainframes? I I won’t even bring up AI yet, but you know, just i in 2026. **Rich Larin (00:55)** Well mainframes are still quietly behind the scenes powering the global economy. Every credit card transaction, airline reservation you make, that’s all handled by a mainframe in the back end. And it’s doing awesome. It is a powerhouse at handling these type of workloads, and so many of our customers depend on them so heavily, and they’ve been around for so long. We’re so excited to have so many new innovations to help handle and help modernize these workloads. **Kate Holterhoff (01:20)** Amazing. Yeah. not quite a mainframe topic, but I I love it. Every time I come to an IBM event, there’s always the the h history exhibit. And so I was I love that, I know. So yeah, I mean IBM obviously has these deep roots around infrastructure and so I I think one of the questions that a lot of us are are grappling with right now is how your history with mainframes is going to be intersecting with some of the new announcements around Bob and AI in general. So how do you see AI and the the larger IBM Z working together? **Rich Larin (01:55)** Absolutely. So the cool things about mainframe is the hardware is as cutting edge as it gets. The z17 is an AI powerhouse, right? It can handle any sort of workload you throw at it from a transactional and security standpoint. But when you look at the applications that are running on mainframes, a lot of them are COBOL, PL1, assembler applications that have been around for decades. And whenever you have something around for that long, it’s hard to maintain, it’s hard to modernize, and it’s almost irrespective of the language. Anything that aged need specialized tools to deal with it. And so that’s where AI is making such a difference to so many of our customers when they look at these applications that have been around forever and they need help being able to decipher what’s the business logic, if I want to make a change or add a new feature, how do I fit that in with this big legacy we have there? And that’s where our tools really are going to be shining with what we’ve been announcing recently with Bob and the Premium Package for Z and some of our other technologies I’d be happy to talk to you guys about. **Kate Holterhoff (02:54)** all right, well let’s dig down into some of that. So during the keynote I heard mention of not only this broader topic of modernization, but COBOL specifically. And I think a lot of us when we think about the mainframe sphere, we we immediately go to COBOL. So talk to me about where that sits. Like w what **Rich Larin (03:12)** yeah, COBOL, I mean, we, our customers, believe it or not, love COBOL. They love it. Because COBOL is such a powerful language for this type of workload. For these high performance, high throughput languages, COBOL is really, really purpose-built for this type of workloads. Okay. The problem is over the years, you haven’t really had modern tools to work with it. Sure. Right? And so when we say, all right, we’ve announced Bob Premium Package for Z, one of the things that really specializes is understanding COBOL, understanding the semantics of the language, being able to make changes to your COBOL applications or integrate your COBOL applications with broader parts of your IT estate outside of the mainframe as well. So that’s one of the most important things around our, you know, our innovation is around bringing the Agentic AI, having specialized tools for this domain that allow customers to almost like treat COBOL like any other language. If you’re using Bob for Java, well it’s the same idea if you’re using Bob for COBOL. If you’re using Bob for Node, like the same type of process and experience is what we’re trying to bring to our clients, so it’s standard. **Kate Holterhoff (04:13)** Yeah, that makes sense to me. Okay, so you are seeing I guess more workloads that are are targeted around modernization and Bob is helping you to achieve success **Rich Larin (04:22)** Yeah, and a lot of our customers for years, like the mainframe was kind of a silo, right? they have their own proprietary processes, these old school legacy tool chains. Yeah. And like we want to totally eliminate that. Wow. An agentic IDE, no problem, give it to the mainframe team. It’s the same one. A Git based pipeline, no problem. Mainframe can participate in that. So it’s the same type of processes that are, you know, with a lot of the announcements and a lot of the capabilities we’ve released over the previous years are now unlocking that to break down those barriers. **Kate Holterhoff (04:49)** Okay. So just in terms of brass tacks, when I think of the Bob interface, I tend to think of a sort of IDE experience, and now that Bob has had its its 2.0 glow up, yeah, so it you know I I I get that the you know the chat box, the natural language way of interacting is is sort of more front and center. What does it look like for for IBM Z? **Rich Larin (05:10)** Yeah, so if you’re a mainframe developer, it is the same exact IDE. The same exact IDE. It’s not a new, here’s the Z version of Bob. But what Bob comes with it when you have the Premium Package, again, same IDE, but it comes with an extra set of features. So if I’m a COBOL developer, I have all COBOL linting, I can do debugging, I can have all these AI interactions that have, you know, let’s say, hey, generate a modification to this COBOL program to add this feature that’s aligned to my existing COBOL standards, right? So that type of workflow that all the developers are loving now that they can do it. and all these agentic IDEs, you could do the same thing for COBOL. I need to generate a unit test for my COBOL program. Bob can help you do that just from the code. So these are all the types of workflows that, I mean you turn back the clock only a couple years, just like it was very foreign. So we’re really trying to encourage our clients to adopt these new technologies, embrace them, and I think it’s going to really help you know sustain the momentum that we’re seeing in the mainframe space. **Kate Holterhoff (06:05)** For sure, for sure. And something at RedMonk that I I’ve heard a lot about is interest in having on-prem and locally hosted models. this it comes from the need to optimize token spend, but also to ensure that these systems are secure. How is it that that IBM Z is helping with that? **Rich Larin (06:22)** Yeah, so it it’s obviously a huge concern for our mainframe customers, so much of their source code they consider IP. but but that being said, we’re seeing a whole spectrum of patterns in terms of how mainframe customers are using this AI. Okay. And the cool thing about Bob and all the premium packages is that you have options, right? You’ve you’ve you’ve heard the announcements, Bob came out as a as a I’ll say like a SaaS only option for inferencing. Yeah. But since then we now have hybrid and fully on prem air gapped options. So if you want air gapped on prem, So can we have certain open weight models that we support that you can deploy on your own GPUs. No one’s ever going to have access to them except you and you can have the same, you know, the same experience from a Bob standpoint. We have a hybrid option where you want to have some part of your experience self-hosted, and then maybe you already have an existing agreement with a third-party vendor, you know, one of the other model providers, but you want to use that, you know, it’s a private, you know, a private deployment of a model that an enterprise has an agreement with. Well, if you want to use that for inferencing with Bob, you can do that too. So we give options and you know I’d say like we’ll see how things play out over the coming quarters, right? We the on-prem release just happened at the end of last month, and so we have a number of customers that are in the process of deploying that. So, you know, we’re gonna we’re gonna see a mix, I would say. Yeah. But I think you know, depending on the customer, the regulatory environment, you know, they they have the right option that they can pick for them. **Kate Holterhoff (07:44)** hybrid workload is the one that I’ve been hearing the most about. So I I I find that your focus on that is particularly useful. Are you seeing that that’s the one that’s the most in demand? **Rich Larin (07:53)** I I mean I think that again, I’d say the demand signals are probably that’s the where the biggest signals are. I think, you know, a lot of these deployments play out over time. And so it’s hard to give you like a a concrete answer like, yep, it’s it’s leading the pack now in terms of it. But I’d say from a signal standpoint, you know, based on where we see, you know, deployment initiatives happening, it’s it’s up there for sure. **Kate Holterhoff (08:13)** Awesome. And I want to drill down on the security question, because you mentioned that that’s obviously the part that many people are concerned about. Anyone who’s following the news recently has seen that security breaches are everywhere. Certainly something I’ve been writing about a fair bit for RedMonk.. So I mean talk to me about how you are approaching this new security landscape. **Rich Larin (08:35)** For sure. So, yeah, I mean, security is almost like criteria number one for when you’re doing this. I would say I’ll talk about a unique thing that we do in the Z space. Right? And so there’s known best practices from a security standpoint of how you want your code to be, whether you’re writing a COBOL or PL1. There’s rules you need to follow to maintain proper security posture. And so what a capability that we have in Bob with the Premium Package for Z is that whenever you make a change to your Code, we will trigger security scan. And it’ll, it’s a capability called Z code scan that’s built into the Premium Package for Z. And so the scan will kick off, and it’s a deterministic tool. So it’s not actually AI doing it, it’s deterministic to see if you have any violations to known security best practices in COBOL and PL1. And so you’ll get your report, you’ll be like, all right, you have these, this is a blocker, this is a high severity, medium severity, you know. And in the old world, like, all right, now a developer’s got to go remediate it. Now Bob can fix all those for you. And so Bob has the knowledge of Of your COBOL practices, your COBOL applications, it sees the outputs of the scan and then remediates it for you automatically. And so that’s one of, I think, the most powerful things around kind of like shifting left the whole secure, you know, I’ll say like your security posture. So it’s got to be in the code from day one. You don’t want to react to that. And so I think that’s one of the most important tools that we’re giving our clients. **Kate Holterhoff (09:55)** Makes sense And I come from an education background. You know, you mentioned the fact that mainframe expertise is hard to come by. and one of the things that makes Bob unique among many is that it you’re able to query Bob about Bob, which is is great. So how is it that you’re filling this knowledge gap by way of Bob for the mainframe specifically? **Rich Larin (10:17)** So we have, so I think there’s there’s a couple things that are really important here. One is Bob and the Premium Package for Z lets new talent coming into the mainframe space get up to speed way faster than ever before. It used to be years where a developer gets hired, and you don’t have to learn, not only learn COBOL, we to learn like every keyboard shortcut to navigate every panel in ISPF. And like, you know, time to proficiency was really long. Now with Bob and the Premium Package for Z, you you’re going from years to like months. And so that’s a huge thing for me. talent standpoint. The other side of the equation is like, you know, you have some of your biggest experts retiring or or or or leaving the business. How do you kind of like preserve that knowledge? And so one of the things that we do is we give ways for, like I’ll give you an example of a tool in Bob, Premium Package for Z called the data dictionary. Like all these like COBOL variables and module names are all like super cryptic. It’ll be like, you know, ABCD1234, and you’re like, does that have to do with customers, accounts? I have no idea. No, no, no. But like, you know, Someone in the back who’s been there for 30 years, of course they know they can rattle off exactly what that is. And so a data dictionary allows enterprises to kind of map kind of a technical concept to a business meeting. And so we have a whole workflow that’ll generate that for you, it can be validated by your experts, and then all of a sudden now Bob has that knowledge where previously had to ask that one expert, now it’s built into your workflows. So whenever you see ABCD1234, you know, okay, that has to do with account management or something like that. And so these types, I mean, we know skills is a huge, huge, important priority topic. our mainframe customers so we’re really trying to be conscientious about having capabilities to to support that journey. **Kate Holterhoff (11:50)** that’s fantastic. I love that. lot of developers right now are concerned about the future of the industry. This is a very hopeful message here that, there’s there’s room in the mainframe space. Come work with IBM Z, this is something we can help get you onboarded quickly. **Rich Larin (12:03)** Absolutely, we are putting so much effort on attracting that next generation of mainframers. We have skills programs, and Bob actually helps these programs be more attractive now. Because now you have, again, it’s not a special mainframe tool that this skill won’t be transferable to future career aspirations. It’s the same type of agentic tech that is going to be relevant anywhere in your career. And so, you know, I’m excited. **Kate Holterhoff (12:25)** I bet, yeah. Well, I mean, hardware is very cool now. I mean, not only do you have the archive back behind us, but also NVIDIA was on stage during the keynote, we’re talking chips, it’s on everybody’s mind these days. Well that’s very exciting. okay, well, as we wrap up this conversation, I I’m interested in any sort of calls to action that you might have. Like what is it that everyone should know, maybe even something about the roadmap? You know, what’s what what should we be paying attention to around? IBM Z. Yeah. **Rich Larin (12:55)** Yeah, so I’d say that I’ll I’ll answer both questions. The call to action is don’t be afraid to try. Get in there, get in the mix. I mean, so often you know mainframers are a little hesitant to try the shiny new tool. Yeah. But you want to be a part of this wave, this era. And so, you know, do a POC, kick the tires, and you’ll be pleasantly surprised. And that that that’s what we’re seeing. And because from like a roadmap perspective, I think one of the most important things that we are doing in the Premium Package for Z is we’re combining AI with the ability to handle super huge applications. And we are have a lot of big releases coming up to be able to couple application analysis and feeding that insight into an AI to give really, I’d say even more impactful answers than you can get today. **Kate Holterhoff (13:41)** man, can you give us any teasers about those big announcements? We’re just gonna have to wait and see. **Rich Larin (13:45)** You’ll have to wait and see. You’ll have to wait and see. **Kate Holterhoff (13:47)** Fair enough, fair enough. All right, well again, my name is Kate Holterhoff. I’m a senior analyst at RedMonk. here live at TechXchange 2026, IBM’s Premier Developer Conference. I’m with Rich Larin talking IBM Z and Bob. And thanks for watching. **Rich Larin (14:03)** Thank you so much for having me.