Name: Skyla Loomis, IBM
Uploaded: 2025-04-08T14:00:00.000Z
Duration: 21 min 28 s

Clips
More from IBM Z: Hybrid Cloud with Advanced AI Where it Matters Most

Skyla Loomis

General Manager, IBM Z Software

IBM

play_circle_outline Skyla Loomis of IBM Z Software: Portfolio, Strategy, and Mission for Hybrid Cloud and AI Empowerment

play_circle_outline Simplification through automation, infrastructure as code, and AI assistance

play_circle_outline App dev evolution with common tool chain integration and AI for understanding complexity

play_circle_outline Core runtime value proposition leveraging data gravity and AI for analysis

play_circle_outline IBM's broader agentic framework and partnership with research for inter-agent communication

23:39

18:51

Info

Skyla Loomis, IBM

Skyla Loomis

General Manager, IBM Z Software IBM

>> Hi everybody. We're back at One Madison for the IBM Z Launch here with Skyla Loomis who's the GM of IBM Z software. We heard earlier from Ross Mauri about the Z platform and for years, when it comes to mainframes, Skyla, we talk about exploitation of the hardware. That's where software comes in. Welcome back to theCUBE. Good to see you.

Skyla Loomis

>> Thank you. Good to see you too. Thanks for having me.

>> You're very welcome. So talk about your mission and the scope of your responsibility and the portfolio.

Skyla Loomis

>> Yeah, so I'm the general manager for IBM Z software and I lead, I like to say most everything above the operating system on the Z OS platform. So our core runtimes, our application and data, but then as well as our compilers and our languages, some of our security capabilities, what we do for developers, but then also what we do for our IT operations and AI ops. So really kind of the whole span of the experience that you really have around the Z platform.

>> How would you describe the IBM Z software strategy?

Skyla Loomis

>> So I think specifically what we're trying to do on Z is we're really trying to remove the friction around the platform. So we want to seamlessly integrate into the hybrid cloud. We want to empower and embed AI around all of that to really dramatically simplify and reduce the skills that are needed to be successful and get as much value out of the platform for our clients as possible.

>> Right. So you're trying to take away everything, we talk about this all the time in AI, take away the mundane tasks that nobody, we saw this with things like with RPA. At first, people were maybe a little concerned and then they were embracing it and thankful. I know in our own experience that's been the case, but I want to talk about, follow up on that and talk about the simplification. I thought the answer you gave around a question that I asked around personas, your thinking is they'll evolve. So how will AI simplify client operations and how will the personas evolve? What should we be thinking about in the next two, three, four, five years?

Skyla Loomis

>> So I think particularly on the operations side, there's a lot of, I think manual work that clients have to do. There's skills that are very proprietary to the platform, JCL, REXX, and so some of the things we're doing is we're bringing enterprise standard automation. Ansible's already been there for a long time. Terraform we're bringing to the platform as well. And JCL is one of those things that job control language, it's a very mainframe-specific type of almost semi program that you write to sort of do things on the platform, to make changes on it. And it's a very unique thing. It's not necessarily that easy to learn. And by leveraging these common enterprise skills, you're able to enable and empower more people to work on the platform and be successful with the platform. And it also I think starts to raise what these personas are doing versus manually running these things, running these jobs or scheduling them, but you're now more starting to move into infrastructure as code, which we've seen happen, that transformation happen to infrastructure folks and other platforms. I think we're going to see that same type of transformation on Z. but then also with the advent of AI, it'll help with generating some of that automation perhaps for you. It'll also make it possible so that it can be just interacting with the chat and you're asking questions, but then relevant automation will come up and be part of offered to you as, "Hey, let me just go run this." Right? It's already been vetted, it's already trusted as part of the enterprise automation. And so we're really, I think going to make those personas able to be much more productive, much more efficient, and I think much more approachable to folks who don't have to have as much deep Z OS core knowledge so that then they can focus maybe on higher level things like how are we optimizing the environment? How are we making it as performant as possible? How are we making sure we've got the fully resilient architecture? So it might become a little bit more systems level thinking as well in addition to just the ongoing management, they'll become more and more automated and AI assisted.

>> Interesting. Take JCL as an example. That was a coveted skillset, and now you're saying, "Hey, let the machine do that for you. You focus maybe on optimizing the platform so we can get more business value and drive value for the client's clients."

Skyla Loomis

>> Right. Right, exactly.

>> You talked about three core, I call them pillars, I don't know if you use that term, but app dev, which drives agility, and then the whole app and data management, that core runtime, CICS is an example, and then IT ops, which is all about being more efficient. I wonder if we could dig into each of those and get your perspective, how you're thinking about the evolution of those and what that means for customers. I mean, obviously app dev, you mentioned dev ops practices and infrastructure as code driving more agility. How do you see that evolving and where does the Z software portfolio fit?

Skyla Loomis

>> Yeah, I mean, I think app dev is really interesting because there's a massive amount of code on the mainframe that's been developed over many decades in most of our clients. And it's all of their business logic. All of their core business logic is built into this, but often they don't understand it. They may have a lot of risk potentially around this and particularly in some of the regulated industries that can be, start to become a challenge. And so I often think of these applications as big tangled up balls of string, and you're afraid to pull on one of the strings because you don't know what it's connected to and what's going to happen to everything else that's related to it. And so what we're really trying to do is make it possible for clients to be able to make these changes and have the right level of business agility with confidence and low risk. And so we're doing that in a few ways. We've been on a journey for a while where we're bringing just common developer tool chain to the Z OS platform. VS Code, Git, Jenkins, Artifactory, I mean, you name it. Basically you can use all of those common tools on your core ZOS applications. So this reduces the learning curve just from a tool's perspective. Developers have all of their creature comforts. Now maybe they're still working on a COBOL application, but honestly COBOL isn't that hard to learn and developers learn new languages all the time. When Go came out, people didn't say, "Oh my gosh, that's a new language, I can't possibly learn that," right? No, they dug it and they learn it. So the problem isn't the language and the problem isn't the tool chain anymore. The challenge is just this complexity of the application and the lack of understanding around it. And this is where I think AI becomes a game changer for us, because we're able to give you X-ray vision into that ball of string. And so when you pull on something, you know exactly what it's connected to. And we can actually provide, we've had graphical ways of doing this, now we're actually starting to have chat ways of just being able to ask questions. Where is a particular set of business logic? If I need to update a motor policy, what transaction actually is that motor policy living in? Where is that data stored? What is the logic to do that update? And it'll kind of zoom you in, finding that needle in the haystack in your application so you don't have to have this deep SME knowledge of your application anymore to be productive and effective and confident than in the changes that you're making.

>> Interesting. So you can visualize those dependencies. When you mentioned all the different sort of tool chains, is that integration or is that an abstraction layer, like a translation layer that you've built?

Skyla Loomis

>> I mean, we've enabled them to just natively understand COBOL, PL1, Assembler so that you can just use all of those tool chains as is. In some cases we had to, for a build to integrate into CI-CD, we have a module that helps do the build integration. But in VS Code for example, we've natively built in understanding of those languages in the debugger, et cetera. We just have plugins that know how to do that. And you've got the exact same experience there as with any other language.

>> And the app dev piece to me is so interesting. And one of the things, the frustrations I hear from that community is when infrastructure is called, sounds great, dev ops sounds great, but they're being asked to now secure the infrastructure. Is that different in Z because of the security ethos? It's sort of designed in and is maybe less onerous for the application developer, or is there a similar dynamic from the security standpoint?

Skyla Loomis

>> Yeah, I think Z is, we consider it the most secure-able platform around. And so a lot of it is built in. We also have a whole range actually of security tools that not only help you to do the securing, but help make it easier to understand what you should and what best practices are there for it. And frankly, these applications have been around for a long time. They've gone through a lot of security, they've gone through a lot of compliance. They're already secure pretty much. And so you're able to take advantage of all that investment over the time and be able to continue to leverage it.

>> All right, the middle layer of the sandwich if you will, the app and data management, so the core runtime, what's the value prop there? What are your objectives? What are you trying to optimize for clients?

Skyla Loomis

>> Yeah, so there, I mean, that's really, I've had a client tell me that that core part of their business is the heartbeat of their enterprise. All of the money flows through it if you're a bank. All of your insurance claims flow through it if you're an insurance company, et cetera. And so you almost aren't an enterprise without those things. And so I think there, we're really trying to help take advantage of the data that you've got and the data gravity that exists on the platform. Again, it's there, it's secured already, it's rich. And when you start to marry what now the hardware is enabling us to do, we have capabilities where we can just within DB2 Z, be able to do similarity analysis of both structured now, and in the future, we're going to be adding unstructured data that may exist in there so that you can actually do analysis of say, insurance claims that are there. And be able to pull out and summarize what the cause was, what the claim is about, be able to assess risk, be able to do a similarity of these claims to maybe fraudulent claims so you can actually determine what the risk factor is on a claim for fraud. And then that's where you can start to bring human in the loop, but you're driving a whole lot of efficiency into your business process because you said, "Hey, a bunch of these are good, and we've got patterns and we know what the recommended approach should be for how we respond to these claims. But you know what? These look a little sketchy." You might want to go examine these, go ask for more information. But we can automate a lot of that process leveraging the AI that we have built into the platform, the hardware capabilities, and then taking advantage and offering that on top of this goldmine of data that we really have available.

>> And this is where again, we talk about bringing AI to the data, organizations have that proprietary data. That's not going to go up into the cloud or to get trained on the internet. That's unique advantage for those organizations, and that's I think something different. I'd love to get your perspective on this then, so the cloud era, okay, I've got my apps on-prem, I'm going to maintain them. If there's compliance things, I'm going to comply, but I'm investing a lot in this sort of new apps. Today, I need to invest in those. You mentioned data gravity. First time I've heard that term today, but it's a really important concept. People, they're not going to move their claims system and these core transaction systems into the cloud. It's too expensive to move. It doesn't make a business case to do so. And as a result, that feels to me anyway, it's ripe for innovation and investment within organizations because the business case is there, the payback and that proprietary advantage, whoever gets there first is going to confer advantage.

Skyla Loomis

>> Yeah. And I think too related to that, I mean, we kind of talked about insurance example, but you think about all the financial transactions that happen and there's like how many hundreds or thousands of ways that people come up with to do fraud right around transactions. And there, it's so important that you catch it in the moment. That timeliness is essential, right? You catch it six hours, 24 hours in a batch window later on, that money's gone. Trying to get it is almost not worth the effort even at that point. But today, often banks will only run their AI models to try to catch fraud on 10% of their transactions. It means 90% are just going through and maybe they'll catch it later. And so there's tremendous opportunity even just leveraging some of the models and capabilities that clients already have and just running them all the time. Because now if you run them on platform with these chips with the performance that we've built into the Z17, that's incredible value and very easy to do.

>> And that's real-time inference, fraud detection is an obvious one. Anti-money laundering, KYC, Know Your Customer, all of those and many more are going to benefit from that real-time interaction.

Skyla Loomis

>> Absolutely.

>> And then IT ops is sort of the third leg of the stool, and that's really all around just efficiency, machines doing things that humans could do, but machines probably can do them better.

Skyla Loomis

>> Yeah, honestly, I am really excited about this area because I think it's one that hasn't gotten as much sort of love and attention over the last few years. And really it's a big data problem effectively, right? I mean, the amount of data that these machines and these applications produce just around how they're running and what's happening is massive. And that's partly why it's so hard and required so much skill, because you, humans had to pick through and try to find things, things, and you would get a little bit of assistance from the machines to do that. But now I think we can take logarithmic levels of improvement in terms of how much we can help our users be able to proactively find the anomalies, correlate all the data together across all these different parts of the system, be able to do that advanced anomaly detection using machine learning, bringing in more AI models, and then actually not just doing the analysis. Because even that I feel like we've had some of the capabilities around for a bit, but now with agentic AI, these reasoning kind self-encapsulated agents and the fact that they can talk to each other, they can know, oh hey, I've got all these tools at my disposal. Oh, we found this as a CICS problem? Let me go to my CICS agent that really understands CICS and can help me to pull the data that's relevant to this. Let me get the DB2 data, let me correlate this together. Now let me give you a recommendation on what the problem might be and how you might start to solve it. And I think we're not quite there yet, but where we want to get to also is providing the explanation of everything that happened. So there's transparency around it, not only so you can trust it so that we're actually teaching our users at the same time. So we're not just helping them catch their fish, but they're teaching them how to fish so they understand what happened and what's going on behind the scenes as well.

>> That's kind of always been the mainframe way. I mean, you think about RACF, who did what, when, where, bringing that to AI, I want to actually jump ahead, stay on agents for a moment. We were talking to Ross earlier and I asked them to difference between assistants and agents and assistants, kind of the uber ex-officio, if you will, an agent's really good at a particular task. What's your strategy around agentic and the framework that you're using to think about it? Because it is early days.

Skyla Loomis

>> So we're really, I think, building on the shoulders of the broader IBM software and research platform. So it's not just, you know, IBM Z in the mainframe that's focusing on agentic. IBM as a whole is investing in this area significantly, and Watsonx Orchestrate is one of the products that is really kind of in the leading edge in partnership with research, building out an entire kind of agentic framework that will be able to be applied to multiple different domains. Also, thinking about how we can help clients build their own agents to be able to kind of plug in. And so they're really building all of that out so that they can have inter-agent communication. How do we define what each agent is so that it knows how to call it and find it? So all of that is really being built at an IBM level through software and research. And then we're really, I think of us as kind of partner zero or client zero within IBM, leveraging it and specializing our application of that to the mainframe with the specific agents, domain-specific agents that then we're bringing to the table that will leverage that framework and then that we can provide feedback and help improve the framework and work together in that way.

>> And one of the big values of course that you bring is that you've got transactions. And we've said if you're going to build a real-time digital representation of your business, digital twin, people talking about them all the time, and these are systems of agency, you can't do that without transactions. That's what transactions enable. I want to close by talking about the business case. I know you guys do some business value assessment. I've talked to many clients, every time there's a new Z that comes out, they say, "We just buy it site unseen because it prints money for us." So I have no doubt you've got many, many orders probably already.

Skyla Loomis

>> Let's hope.

>> I'm confident. My question is, two part, the core business value, and how does that core business value get enhanced with AI? How should we think about that?

Skyla Loomis

>> Yeah, I mean, I think at the end of the day, we're really trying from the software perspective, we're really trying to drive productivity, efficiency. And I would also say risk reduction is the other thing, which is maybe a little bit hard to measure, but I think on the productivity side, we are focused a lot on, okay, you had to, as a developer maybe from either a low end of experience to a high end of experience, go in and try to understand where in this application that motor policy lived, how long would it take them to do that? They would've been literally searching through code or the newbie would've, right? Maybe someone experienced says, "Oh, I've worked on that a few years ago. I think maybe it's over here," and they might have a little bit starter period. But now with the tooling that we provide, we can have people come in pretty much cold and be able to see how long it takes them to be able to pinpoint and identify the same thing. And we're seeing two to 5X productivity improvements. We're seeing Sun Life actually as a public reference we have where they were seeing a 50% productivity-

>> 5-0?

Skyla Loomis

>> 5-0 in terms of finding the code and understanding the complex relationships that was there. So I think we can look at each stage of what happens in a developer life cycle going from trying to understand what they need to do, actually doing the code changes of maybe that refactoring. And especially now as we're starting to add in more smarts around what each of these things are, actually building a business glossary behind the scenes so we can understand the elements of the application so you can just from a simple chat, be able to quickly ask these questions, and again, pinpoint you and bring you right down to that part so we can start to measure where we see those things. We had one company we were working with, a US global logistics company, and they were, I guess global logistics company I should say. And they were actually looking at how do they modernize and transform and really re-architecture their applications from co-procedural COBOL right into object-oriented Java. And we help with that use case as well. And we were to leverage our IBM consulting team to help work with them on that. And again, these aren't folks who necessarily know the applications as well, but they come in and they were in two months with about 12 developers able to transform nearly 400,000 of code. I mean, I think that's kind of levels of, and this isn't just kind of push a button, machine, in machine out, you gorp in, gorp out, right? This is true. Let's re-architect. Let's make this subject-oriented. Let's get to high quality output of the code.

>> Yeah, lots of software innovation, lots of value that you guys are driving. Thanks so much for spending some time with us, Skyla. Appreciate it.

Skyla Loomis

>> Yeah, thank you.

>> Okay, I'll be back right after this short break to wrap up. We're at One Madison for the Z launch. Keep it right there.

Skyla Loomis, IBM

Skyla Loomis

>> Thank you. Good to see you too. Thanks for having me.

>> You're very welcome. So talk about your mission and the scope of your responsibility and the portfolio.

Skyla Loomis

>> How would you describe the IBM Z software strategy?

Skyla Loomis

>> Right. Right, exactly.

Skyla Loomis

>> And that's real-time inference, fraud detection is an obvious one. Anti-money laundering, KYC, Know Your Customer, all of those and many more are going to benefit from that real-time interaction.

Skyla Loomis

>> Absolutely.

>> And then IT ops is sort of the third leg of the stool, and that's really all around just efficiency, machines doing things that humans could do, but machines probably can do them better.

Skyla Loomis

>> Let's hope.

>> I'm confident. My question is, two part, the core business value, and how does that core business value get enhanced with AI? How should we think about that?

Skyla Loomis

>> 5-0?

Skyla Loomis

>> Yeah, lots of software innovation, lots of value that you guys are driving. Thanks so much for spending some time with us, Skyla. Appreciate it.

Skyla Loomis

>> Yeah, thank you.

>> Okay, I'll be back right after this short break to wrap up. We're at One Madison for the Z launch. Keep it right there.