BMC Enhances Mainframe App Development to Meet Modern Enterprise Demands

Hey guys. Thanks for the throw. We're here with John McKenney, who's executive vice president and general manager for the mainframe business unit at BMC.

And we're talking about a series of updates they've made to the tools they provide that make it simpler to build mainframe applications. And there's also a little bit of a study that they did with Forrester about what developers actually need. John, welcome to show.

Thank you. Good to see you, Mike. Yeah.

Well bring us up to speed that I realize that there's been a series updates including an alliance with Okta, but what have you guys been doing and, and how does it change the way we think about building these applications? Well, you know, I think, uh, we've chatted a lot in the past and we're still very focused on, you know, helping make developers and everyone that works on the mainframe platform more productive. So we think the best way to do that is to bring the, the latest and greatest technology that, uh, many of us are able to use.

And, uh, some of the things that we're doing, we're not working and bring that into, uh, the modern mainframe workspace. So for application developers, you know, we're making it easier for them to understand, develop, and create code. Um, and some of the latest capabilities that we have, enable them to work right in that BS code interface, get a chat assistant kind of, uh, help to understand what the program is they may be working on or looking at for the first time.

And to be able to confidently make, uh, the needed changes to that program to deliver some new business service or new capability. So, uh, that's, uh, that's one of the things that, uh, we continue to make progress on each and every quarter. You guys have been working on this for quite some time, and now we've entered the age of ai, and I wonder if AI will make it easier to build these applications.

'cause historically, you know, they were really the realm of a particular specialist. Yeah, I think, uh, you know, the power of generative AI is that it can really synthesize and summarize some really complex concepts. Uh, you know, it does that by, you know, creating, uh, some new content.

So there's a couple of ways, right? So one is, as I said, you can help understand an application that's written in a similar, which is a really lower level language, uh, or understand an application written in PL one or Cobalt. So if I understand the application better and more quickly, then I'm able to make the necessary changes I have to it.

But the next step is obviously to be able to help me make the necessary change I want. But it's funny, when you talk to application developers, you know, if you're not a developer, you might think, oh, all they do is code. Well, there's a lot more that a developer has to do, then code, you know, she or he may need to actually test that code.

So we can help with test data generation to support that test. We can help with creating the test. So these are all things that, you know, BMC has created tools to help developers do those other things that they must do to get a new application into production.

And we believe with a assistant, you know, we can make those jobs even easier for that developer. So what does it mean for their, you know, them as a developer, they're more productive. They can, they can get more new capabilities, new features out with the same or better quality and and confidence than they could before, which is, uh, you know, a good thing for them and a good thing for their, the business that they work in.

Now, you also recently did a study with Forrester looking at the developer community itself and how it was evolving. So why don't you give us some of the highlights of that and what surprised you of anything in there? Yeah, you know, I think, uh, some of the highlights from the Forrester study, you know, really, um, underscored as I was saying, you know, the things that developers spend their time on, um, for many organizations in the mainframe, um, you know, there's, there's a change in the workforce, right?

We, we have people that are retiring and new team members coming on board. So if you're a new developer, uh, one of the things that the Forrester study found is if you are using our B-M-C-M-E DevX solutions that the clients using, those reported that their developer onboarding time was cut in half, so from about nine months to about four and a half, right? That means that that developer, uh, can be productive for their organization in a much more rapid timeframe.

Uh, they also reported that the, uh, the ability and speed by which they can get new capabilities and new features out into production is shortened considerably. Um, and so these are the kinds of, uh, tangible benefits that anyone that's looking at making changes, uh, changing some of the processes they may be using and the tools that they may be using, what are the benefits? And there's some very, you know, strong tangible benefits that the Forester study uncovered as a result of, uh, clients that have shifted and began adopting and using the BMC Amy Devex solutions.

Right? Um, it seems to me at least that a lot of the younger developers still are, you know, they want to be where the cool kids are hanging out and they perceive that that is, you know, the next great programming language or whatever it may be. But if you go talk to the people who run companies, especially larger ones, they're still keen buying folks who know how to build and deploy applications on mainframe.

So how do we kind of bridge this divide? How do we get that next generation kind of excited about the mainframe? You know, I, I think there's a lot of, um, you know, early in career, um, technicians out there, developers out there that are very excited to work on the mainframe platform.

I, I get the opportunity as I meet with customers around the world to meet with many of them, and I ask them, you know, why did, why did you guys choose this as where you wanted to work? And, you know, there's some common qualities there and some common themes that I hear. So one is, you know, they really think it's exciting and interesting to work in an organization that has some truly mission critical applications that support every one of us every day.

So if they're in a financial institution, you know, I'm really supporting the core banking system or the loans applications, and these are, these are things that we all use, or I'm part of the payment processing and, and who doesn't use some form of payment to buy a coffee or a meal or go to a movie, or whatever it might be. And so they, they have, you know, an interest and a desire to work in those types of environments where they know, Hey, these things have to work all the time. So that's kinda really exciting for them.

But the other thing they'll say is, you know, I've done my work and I've looked around and I see there's a lot of really new technology that's available in the mainframe. So it's not like, uh, some of the things that they may have or other people have read about in the past that the mainframe is old and it's legacy because it's not, the hardware is very, very modern. It's, is, it's modern and, and, uh, up to date and more current and more securable than other platforms.

And the tools that are available are frankly the same kind of tools. You can work on other platforms, and it supports multiple languages. By the way, most developers, you know, they don't think of themselves as only writing one language.

Most of 'em write in multiple languages, and they find it kind of fun to pick up another language. And so if, uh, the team they're joining writes in COBOL or PL one, Hey, I can pick that one up as well, and maybe I can introduce my teammate as some of the other languages that I've, I picked up is, uh, I started my career and there's an interest in appetite in more organizations to take advantage of that. Mm-hmm.

I also feel the applications are becoming more distributed in the sense that the mainframe is part of the platforms used to build these things. And is that changing the way we think about the types of developers and the skills they need? Because I'm gonna build, I don't know, maybe the analytics part is running some other server, and then the transaction part is running on the mainframe.

Does that change the dynamics? Yeah, I, I think that, uh, it does to an extent, right? So if I think about, you know, when, when I started my career many years ago, you know, an application, you know, you, you wrote all of it in the same language, uh, because you were working with the old 32 70 green screen that, uh, was direct connected more or less to a mainframe backend.

Well, well, today it's, it's very different from that, right? We all operate, uh, and access many applications from our mobile devices, right? I mean, it might be our watch, it may, may be our phone, it may be, you know, whatever.

And I, I kind of think of it this way. Applications today are a little bit like, you know, a home, right? If you got an apartment or a building or a house, whatever it might be, and you think about how that was constructed, and there are many different skilled people behind that, right?

Professionals that may have been the architect that designed the overall thing. You might have, uh, the, the person that was the plumber that put in the, the plumbing or the hvac, the electrician. Then you've got the finished people, then you have the interior work.

So there wasn't one person that did all that. And you think about today's applications that translates into the developers. You've got developers that specialize in the front end applications that we're working on from our mobile app.

You've got others that are responsible for that middleware trans transaction layer and the infrastructure layer. So it, chances are, when we're thinking about your favorite mobile app that's accessing a core business application, there's probably a number of different individuals with different skill sets that are working together to deliver that capability that we enjoy each and every day. Mm-hmm.

Of course, we are all living now in the age of AgTech ai. And as these agents come along, how much will I need to know about the underlying programming languages? Or will I just ask the AI agent to go do something and it will know how to program in these things, and they'll talk amongst themselves about how to build a distributed app.

Yeah. You know, I think like a lot of things, I think there's gonna be an evolution, right? So, you know, today, um, you know, developers can access generative AI to get explanations, to get code recommendations and, and even generate code.

So one thing the developer of tomorrow is gonna need to know and understand is what are the business requirements, right? They're going to be able to describe in some detail what, what they want. Um, and they might leverage generative AI to create that.

Well, most of us, you know, most business applications aren't created in isolation. They've got to integrate with other systems, and that makes some of what generative AI has to do, maybe a little bit more complex. Um, but I do think that developers of tomorrow will get great aid from generative ai.

They'll still need to review it, still need to make sure that it is doing what it was intended, what they wanted in a prop. And we all probably have experience with that, where we've asked gen AI to do something and it didn't quite turn out like we wanted. And so maybe we had to try two or three or four times, and maybe it gave up and just did it the old fashioned way, right?

So I think there's gonna be an evolution, but it will continue to evolve, and I think it'll continue to get more powerful, and I think it will continue to make teams and individuals more productive. You know, I think it's an exciting, uh, exciting future. Yeah.

Um, as that all comes to be, are you seeing organizations do anything in particular that you kind of wish everybody else would cribble a little bit when it comes to AI and how they're adopting it? Because I think everybody's kind, uh, doing it is as a giant experiment, but I'm not sure everybody has a formal plan. Yeah.

I think that I, I see a variety of engagement levels. Uh, you know, at one end of the spectrum, you know, I see organizations that are already developing generative AI policies. We have those within BMC, uh, and guidelines on, you know, where and how we use them, the dos and don'ts, if you will.

And I see other organizations that are doing that. Um, some organizations are going in so far as to, um, being able to, uh, really review the different large language models and small language models and have certain approvals for that. So I think, you know, there will be guidance that will continue to evolve at the other end of the spectrum.

You know, I'd say I still run into some customers where they're not experimenting yet, you know, so my recommendation would be, you know, jump in, uh, begin to experiment, you know, experiment, you know, with some small projects to see and learn how and where ative AI can help your teams. So, as an example, you know, one of the things that, that we have, uh, just, um, made available is our Amy e assistant chat is now available in our A ME ops and our a e Devex products. Uh, so clients, developers or operations personnel in the tools they're using can ask for assistance.

They can ask Amy, uh, for information on how I might do something, and they'll get access to that type of information. So, again, kind of recommendation, step by step, try this, try this. Um, and we will have also in the beta, the ability for those customers to add their own documentation.

Maybe they have some standard operating procedures, so we can adjust that. So within the, the, the, the workflow that a developer or an operations personnel is, uh, experiencing, they can ask for some guidance and assistance. So, you know, I I would say, you know, lean in, experiment with it, try it, see what works for you and your teams, and provide the guidance back to BMC or the other vendors that you're working with.

That's how together we can continue to iterate on technology and make it more powerful so that it really provides the, the capabilities that, you know, clients and, and users want to see. Mm-hmm. Do we have reasonable expectations at this point for ai?

It seems like, you know, I turn around and you'll hear people talking about they're gonna be cutting development teams in half. Uh, that may seem a little, uh, optimistic or pessimistic depending on who's involved. But, um, what is a, a reasonable expectation for the impact AI is gonna have on application development?

Yeah, I think that, you know, a reasonable explanation, you know, expectation, let's stick with the development team. Uh, and they're using generative ai, right? I think a reasonable, you know, expectation is they should be able to see, you know, 10 to 15% productivity improvement initially.

I think over time, that will grow. I think over time, you know, they'll see a bigger impact on how generative AI can make a developer more productive. As I said, it can help them understand code, it can help them change code.

Eventually, it will be able to help them build, uh, the data that they need to test code, build the test cases for it. Um, you know, we're already working in the lab with some, you know, proof of concepts to be able to help guide you through an automated, uh, debugging of a failure and a auto correction and how the developer review that, right? So I think that you're gonna see the, the types of productivity improvements will climb, but I would, I would start with something modest.

You know, let's, let's not expect that it's going to do 80% of, uh, developers, you know, activities. That's not realistic, but 10 to 15% that grows to 20 to 25% that moves up closer to 50%. Yeah, I think that's definitely possible in the foreseeable future.

Moving need, need to change the backend of the way that we build applications. Because right now it seems like most of the focus is on, uh, giving developers AI tools to write more code, but that's only a small percentage of the overall task. Um, will we need to rethink the workflows and the pipelines and everything else that we're doing?

Because as more code is written by AI agents, it seems like maybe the existing workflows will be overwhelmed. Well, I think, you know, some of the workflows around, uh, product development and product deployment, I think they'll continue to evolve. You know, I think the most advanced, have some really good gates in there to, to really judge and make sure that you've got good quality as you're progressing, that you've got the right security in place.

Those will continue to be important. There may be some other things that, uh, you know, go into those types of gates that are trying to understand privacy implications as well as you think about journey of ai. Um, and I expect that, you know, we'll see more workflow changes, you know, around integration.

Uh, you know, that's one of the most important things when you're thinking about the mission critical core applications, is that these things, you know, these applications will perform at scale, you know, with high volume loads. And, you know, those are the things that, again, I think, you know, for the foreseeable future, you know, it's gonna require a lot of human interaction of the developers planning, of the infrastructure teams testing beyond one piece of code, but how all of these pieces of code work together to de, you know, to deploy and provide the capabilities of a, you know, a large complex overall application. Mm-hmm.

And to your point about that, we talk a lot about building of new applications whenever we talk about ai, but there's all these applications that have already been deployed over the last three or four decades. And does AI present an opportunity to go after the modernization of those applications in a way that might be more accessible and affordable? Yeah, I think so.

I, I think that, you know, we have within our Amy Devex tool set, this is, I think one of the reasons the Forrester TEI study showed, uh, productivity improvements for teams that use the tools. You know, one of the capabilities is a tool that, you know, we talked about before called Code Insight. And code Insight can help you understand a very large, maybe older piece of code that some might describe as monolithic.

You know, maybe it's got hundreds of thousands of lines. Well, you, you can understand that and break that down into smaller segments. Uh, you might choose to take some of those segments, uh, and maybe convert them from COBOL into Java or into Python where you might have, feel like you've got more team members that are more familiarity with it.

Or if it's in Java, it can run on a specialty engine so it can be more cost efficient for you. So there's a lot of capabilities there where generative AI can help you do that type of analysis. Let me understand the code, let me understand the different components of the code.

Let me score those components to identify which ones might be really good targets for me to consider to move into subroutines. And as I do that, maybe I, maybe I change the language of that particular subroutine from what it is today into something that's different. Those are all things that, yeah, you and I could do that manually, but it's gonna take us quite a bit of time.

With the right tools, you can speed that effort up considerably and, and get the same or better result, right? Well, folks, you heard it here. There are new tools and AI that are changing the way we think about building and deploying applications.

The only issue now is maybe the limits of our imagination about what to do with it. Hey John, thanks for being on the show. My pleasure, Mike.

Good to talk with you. Yeah. And back to you guys in the studio.