Optimizing Mainframe DevOps and AIOps with Strategic Insights from Generative AI Services | BMC Connect 2024

Hello, welcome. I'm Steven Dickens, and I'm joined by my host Mike Ard. Yep.

We are coming to you from BMC Connect and this is a six five on the road. We're joined by Anthony and my dear friend, Dave Jeffries. Hey guys, welcome to the show.

So let's get started. Tell the viewers and listeners a little bit about what you do for BMC. So, I'm, uh, responsible for really all the aspects of r and d, um, whether it's the, the products that we've had for a number of years or whether it's late and breaking Amy Platform, Amy assistant.

It's a fantastic place to be. So we, we just drive innovation and that's, uh, it's my team worldwide that do that. That's not a bad way to describe your job, Dave driving innovation.

Anthony, you So I'm a architect for Amy platform and Amy assistant and an AI evangelist. Uh, go out and talk to many customers about what We're, is that a cool job title for This year? Yes.

That just appended right onto my, uh, Title. Fantastic. So you took us there, guys in your introductions with generative ai.

We're here at Connect this week, just come off the keynote. Lots of focus on generative AI assistance. Kind of where are you seeing that kind of fit within the DevOps and the sort of space?

I'll go to you first day. Right. So I I we've done some fantastic announcements today.

Mm-Hmm. And, and if you haven't seen them, it's all, all about Amy assistant team platform and bringing generative AI to where we think it's needed the most. Mm-Hmm.

And obviously a lot of people talk about skills, gaps and skills attritions on, on the mainframe. Um, but you have to think, you know, why, why is that an issue? Why is, is the skills challenge an issue on the mainframe?

'cause surely, you know, it's going the way the dodo maybe, or something like that. Well, in reality, it's completely the opposite. The mainframe's got an entirely new lease life, and those skills are probably an inhibitor or the lack of those skills are an inhibitor to transformation.

People want the mainframe to do some new cool stuff because it's got fantastic new technology in there. And we think generative AI is, is providing really that key to unlock what applications do. So therefore, you know, reducing the risk of changing those applications, what your systems do, how your systems into operate, et cetera.

And it's allowing you to, to unleash transformation, which is bringing, you know, a whole new realm of possibilities to the platform and, and how the platform can support business. Bringing new people into the platform. Yeah.

Enabling them to get started faster. Yeah. And it's making, I think the guys who are already there, guys and gals are already there helping them as well to, to innovate because sometimes, you know, you might be the last one there and you might be struggling in terms of, you know, the, the scale of the, of the challenge in front of you.

You need some help to go do it. And I think it's not just unleashing the next generation of talent, it's unleashing the talent that already Exists, Which is important. Exactly.

You touch the customers a lot. Yes. What are some Of the examples that people are actually using here?

Because I think we talk a lot of theory with ai, but you know, what are we actually seeing? Where is, where is the manual effort in the scut work disappearing? Yeah, So, um, it, that's a really good, interesting question.

So from different customers, it means different things to them. So a lot of them it's, uh, capturing that tribal knowledge. Let, let's just start there.

So before they even go down the road of, you know, AI or generative ai, they have to take that step back and see what does it mean to their business. And if you just look at some of the tooling that's out there today, currently when it comes to generative ai, it's very agnostic, right? It really doesn't mean anything specifically to a customer and their wants and their needs.

So the first thing that we had to do is take a step back and say, well, how do we infuse our generative AI with the knowledge from a customer's environment, first of all to make that AI relevant to them to address their wants, their needs, their business direction. 'cause that's where, you know, uh, I think you used the word skunk works. What, what, when it comes to that, well, how do we make it a reality?

How do we make it relevant for the customers? So we built our platform of generative AI services in a way that it's open and customers can infuse it with knowledge that they need and then start applying it to things in the DevOps space, in the, uh, AI ops space specifically around, uh, you know, what they're trying to get out of, uh, improvements in their, in their business with the generative ai, making it relevant in context for them. So I can customize it and it meets me where I am versus me Exactly.

Being forced to do something. That's Right. So you don't want to ever leave your experience or your environment or your tooling to jump out into another, you know, platform or another tooling.

'cause that's where you lose what we call context and the relevance of what it means to you. So we like to say we meet the customers where they are in our product experiences with that context, with the understanding, and you get far more outta generative AI and much better results that are again, relevant for you and your business as you move forward. So Dave, key word there from Anthony was spec specific, making this specific for the particular shop, the operators Yeah.

The developers. Can you just kind of double click on that as a phrase and what that means from a specificity, you know, contextualize it, what am I gonna be actually doing with a assistant or some of the a a AIOps stuff, right? And, and how that's gonna work.

So I think one of the key aspects of all this is, and, and everybody's, you know, they've approached generative ai. There's, there's a world and a plethora of LLMs out there, large language models. And we we're starting to see some language models being really good at certain things in certain areas.

Some are good at codes, some are good at other things, et cetera. So as Anthony was talking about, you know, one of our kind of think core traits that we bring to the, to the platform and to our solution is allow you to take the right LLM for the right use and then obviously infusing it with your own Information. So that's how you're getting that specificity.

Oh, absolutely. And then you can apply it to the code world. Yeah.

We talk a lot about generative AI in code, in terms of understanding code, but there's more than just code that runs the business. There's the infrastructure, there's the configuration, there's the environment. And so not just applying generative AI to understanding what a Coble application does or what a an assembler application does, but what does the recs do?

What does the JCL do? Mm-Hmm. What does, um, what does the, you know, the, the expert back at base do in terms of how he resolves a particular situation that may appear in operations.

Exactly. So bringing that subject matter expert to a wide variety of areas involves multiple kind of tribal knowledge, pockets being pulled together, multiple lms, being able to be used for the right reason and the right purpose at the right time. I think explaining code is great.

'cause a lot of folks, they didn't document it in the first place. Yeah. So they don't really know how it works.

But how do we go to the next level? Because I think what we're moving now towards is real time insights that are gonna be surfaced as I'm trying to perform a task. And Yes.

Yeah. So we're on this journey, but it seems like there's multiple phases. What are they?

There Is. So, uh, the way we are looking at the spectrum right now, and you know, we started off in the area of what we all started to experience, like with chat GPT, right? It was the chat experience.

So we go out there, we, we dump, dump questions over, right? Uh, how to, what is, how can I type questions to chat GPT, that's how we all started. So we also started that way with infusion within our products where it made sense.

But the spectrum now, we're looking much wider, much far beyond just the chat experience moving to towards something called agent or agent AI or AI agents. And what we want to do there is really move towards more autonomy with our generative AI solutions, but also the point of hyper focusing in our particular product areas with AI agents to be super experienced, super knowledgeable, super capable within a given product area like AIOps, SecOps to do things like, uh, the automation. Let's just take automation for, we have a lot of mundane tasks that we deal with day in and day out, right?

If even if you look at us individually, uh, there's a lot we can start looking at for generative AI in our own lives to simplify things, nevermind in the business world. So we have these AI agents now that are focused in our product areas to allow customers to automate mundane tasks, to surface insights automatically. What do we really want to do here?

It's not that we want to make, uh, we want to, we want to take the cognitive load off folks with generative ai. 'cause we want them to focus on more important business issues. We want them to focus on innovation.

They wanna innovate just as much as we do. So how much could we help them with utilizing generative AI to push them to that journey? Anthony, I think you said it well.

We want them to focus on the good stuff. Exactly. Folks, less on the operation stuff.

What a great way to wrap. You've been watching us here on the six five coming to you live from Connect with BMC. I've been your host, Steven Nickens.

Join as always, by my dear friend Mike Ard. Please click and subscribe and check out to the other episodes and we'll see you next time. Thank you very much for watching.