Transforming IT at Hewlett Packard Enterprise through Agentic AI with Brian Gruttadauria, CTO of Hybrid Cloud

(Intro Narrator at 00:00:01) Today we're talking to Brian Guttadauria, CTO of Hybrid Cloud at Hewlett Packard Enterprise, about how GreenLake intelligence is changing hybrid cloud. You're listening to Joel Beasley, Modern CTO.

(Joel Beasley at 00:00:16) You lead AI strategy at HPE. Is that right?

(Brian Guttadauria at 00:00:21) I do a little bit of it all, right? So the official title, you know, CTO of Hybrid Cloud. But I would say 90% of my work of the past year, year and a half, has been all AI related. Part of it was building out the PCAI project. You know, this was our partnership with NVIDIA. We're doing a lot of AI within our products. A lot of it's agentic. Everything's agentic these days. But we're building that directly and deeply into our products right now, leveraging A to A, MCP. And we also have a series of AI projects that I'm leading across HPE to solve our use cases, right? We have a myriad of use cases that we could leverage AI for, and we've delivered quite a few of them already. One was a contract analysis project where we analyzed tens of thousands of GreenLake contracts. We looked for anomalies, differences. And then when that pesky tariff issue came up a few months ago, we also used it to analyze for tariff risk. And it really did a good job. Not only, you know, these are contracts from probably over 20 years in different languages. We were able to synthesize it and pull it down and do the analysis of it. So it was pretty good. What we found is the biggest hindrance to a successful AI program is having the right subject matter experts that understand how to look at a contract, what it means. Because most of the engineers that are doing the development work don't really know the legalese of a contract. So pairing them up, you know, really drove it to be a successful outcome.

(Joel Beasley at 00:02:20) Was that a project you did for a client?

(Brian Guttadauria at 00:02:22) We did it for ourselves. It was actually, the code name of it was called Ask Bob. That's just, not sure why the team came up with that. But we've now since productized it, and it's going to be an NVIDIA blueprint, basically a NIM, that we're going to integrate as part of our PCAI product. So very soon in our next release, it'll be available for customers to leverage as well.

(Joel Beasley at 00:02:55) Bob is probably the one lawyer that speaks multiple languages.

(Brian Guttadauria at 00:03:01) He is.

(Joel Beasley at 00:03:02) That's fun. Okay, you mentioned briefly this GreenLake intelligence. I actually watched a video. And I really liked, you know, we like to be the top tech leadership podcast. We like to have the best people in the world. And I saw that you guys are on that top right magic quadrant for Gartner. You're the leaders in this space. So what is GreenLake?

(Brian Guttadauria at 00:03:27) Okay, so GreenLake itself is a platform in essence. It's designed to deliver an as-a-service experience to our customers with a cloud consumption model. From that platform, we give you the ability to manage, monitor, maintain, and do any operational aspect of your infrastructure, both on-prem and in the cloud, right? And what we've done to build it together from an architecture perspective is we've integrated several technologies from key acquisitions we've had over the past few years. You know, our CloudOps suite consists of observability from OpsRamp, which we've integrated into GreenLake. We acquired Morpheus Data, which gives you your cloud management platform, gives you the ability to build out virtualization instances, container instances, both on-prem and the cloud. We have full FinOps capability of the products. On top of that, we have full-fledged DR functionality with Zerto. So we have high availability built into the product. And then, you know, all of this is linked in with our Alletra storage arrays, our servers that have full operational management of it. And then we also support third-party infrastructure as well. So you can manage not only our infrastructure, but third-party infrastructure leveraging this platform end to end.

(Joel Beasley at 00:05:06) And from what I saw, it's this agentic mesh that monitors everything and then offers you observability or insights and can help you make changes. Where do you, how do you decide when you're building a product like that, where you draw the line, where you put the human in the loop interaction?

(Brian Guttadauria at 00:05:22) Yep. So one of the things that I kicked off at the beginning of last year when, you know, before MCP, the acronym, became as commonplace as it is now, as part of my CTO organization function, I basically challenged all the teams to really investigate how best we could apply agentic AI to managing our infrastructure, right? This is when MCP was just evolving. The agentic mesh, now A to A, has pretty much become the de facto standard that we've built into it. But even prior to that, we were starting to enable some of our devices with native MCP support. So our X10K object storage array, you know, you can use a tool like n8n, connect into it, and directly get to the object store without making direct storage calls to it. You could just use it as an interface via MCP. So once we started seeing the ability of what we could do there, we kind of opened up the lens and the spectrum of what we can manage. We said, why don't we manage all of our devices the same way? Most of our devices enable themselves via APIs, so we, in essence, enabled all of our devices via MCP. We've developed an agent registry. The agent registry gives us the ability to know how we're communicating. We have, you know, we're leveraging a foundational model, and we're also using knowledge graphs and a knowledge model to understand how it should talk to these agents, right? Using the agent card itself, so it knows how to communicate, what to do, the rules behind it. There's a lot of what is evolving as we speak is the security mechanisms around that. So we're working really closely in the industry to help define how security is going to be working with it. Right now, OAuth 2.1 is the API security model we're using for access. We're looking at task-based security models to give you finer levels of granularity and also putting in basically guardrails around the responses that we have. But all of this is really, you know, it's primarily enabled because we have that GreenLake platform that I was mentioning to you a little bit earlier. Without that, you know, you'd have more of a point-to-point type communication. With the GreenLake platform and what we're doing to gather information about your infrastructure with OpsRamp, so we have continual full-stack observability around your network, and this is across all of our products. This is, you know, we have integration into Juniper switches, Aruba switches. So it's the entire platform end to end. We can now make sense of the type of questions you're asking. So for instance, if you wanted to create me an environment with 20 VMs leveraging 20 terabytes of storage on the highest performing VLAN in my infrastructure, it would know this based on the history it has. And what we do typically, because this is such a new technology, you mentioned in the original question, human in the loop, that, you know, we call it HIL, we have an acronym for it. In essence, you know, the early stages of this will always be, this is what we found, this is what we'd like to do, would you like to proceed? We will ask the user, and we're being very transparent, keeping a lot of the history of what we're doing, allowing you to look at the type of decisions the model is actually making. Everything's being logged so you can, you know, it's going to take a while for customers to become comfortable, I'd say, with letting AI run wild in their network. So we've taken the approach that human in the loop is going to be a big precursor for this for the foreseeable future.

(Joel Beasley at 00:09:38) The trust has to be earned, right?

(Brian Guttadauria at 00:09:40) Exactly.

(Joel Beasley at 00:09:41) I'm a Tesla owner, and people that don't have Teslas often ask me, how are you so comfortable with the full self-driving? It's like, well, at first I wasn't. Like, at first I was just like, let's see if it can get me across from this house to that house and just watch it pull out of the driveway and go down the street. I was like, okay. Now let's see if it can get me to the grocery store. And then it, and then somebody did something and it responded how I would have responded. And I was like, whoa. And then, you know, those little things build up over time. And I think that's how it's going to work with these AIs.

(Brian Guttadauria at 00:10:12) So do you now close your eyes and let it drive you to work?

(Joel Beasley at 00:10:15) I would sleep if I could. I have full and complete, it has earned my trust. It has handled situations that I would have bet money it wouldn't have been able to handle.

(Brian Guttadauria at 00:10:27) Really? That's cool.

(Joel Beasley at 00:10:28) Have you spent much time with full self-driving?

(Brian Guttadauria at 00:10:31) I do not have a Tesla yet. It's like, you know, one of those things on my list at some point. You know, most of my driving is from here to the airport, so it's probably not very useful for me.

(Joel Beasley at 00:10:41) Yeah, that's what I do. I just hit it. You don't have to match the Google Maps with the off-ramp if you live in a big city. It just knows. I'm like, I'm sold, you know.

(Brian Guttadauria at 00:10:54) I'm going to have to get one.

(Joel Beasley at 00:10:55) You should at least try it. Just do a test drive. But, people have told me about those for years, and I didn't get mine until December. So I've had it almost a year. And I was pretty impressed, but more so from the analogy of how I think we're going to be adopting AI where it's like, okay, I'll give it a chance. Yeah, let it pre-configure, let it show me what it's going to generate and see that, okay, cool. And then you do that and then after a time, you're like, it gets it right all the time. Like, why am I even checking this? Just autopilot, you know?

(Brian Guttadauria at 00:11:24) We see the same thing with, you know, we've instituted the use of Copilot, GitHub Copilot for our developers as well. And we see that same kind of paradigm. Like, we see the senior developers accepting less code because they understand that some of the code that's being returned is probably not worth adding in, but more, you know, junior developers blindly accept the results, which is probably a bad thing. So as this evolves to your point, we're going to have to find that happy medium where junior developers who may accept things, how do we inform them or log it or do the change management so if they do do something wrong, we can roll it back, right?

(Joel Beasley at 00:12:10) Yeah. Being able to roll it back. And to be honest with you, I love it because I've been at this for over 20 years. And so it just makes my experience more expensive. As I'm using these tools, I can see how juniors would just be like, oh, that sounds right, that sounds right. Like, let's go. But if you actually run production applications, you know, in high stakes environments before these tools existed, you really understand the areas where things can go wrong.

(Brian Guttadauria at 00:12:35) Exactly. Exactly.

(Joel Beasley at 00:12:37) Before we continue on in the conversation, you did mention that you were doing a talk in Paris, a little bit of fintech, a little bit of healthcare. Now between you and me and a bunch of other people, but between us, is there really much of a difference in those talks or is it just like for those specific industries, you just kind of tailor it a little bit because they're both highly regulated and very complicated?

(Brian Guttadauria at 00:12:59) The talking points are fairly similar, but the, and I would say the challenges are fairly similar across the industry. You know, in the talk we're going to be having next week, they're going to bring it in customer specifically. It's going to be HPE and NVIDIA from the perspective of, you know, we're the arms dealers, we're the vendors that have solutions, and then they're going to bring in some customers that have their pain points that they're going to discuss. But in most cases, it's a highly regulated industry. Most of these companies are still yet to define a strategy in regards to how they're going to apply AI. Now, like, in the healthcare space, there's some point solutions that exist today. You might have medical record imaging. You might have dictation where the doctor's dictating the procedure he's going to perform and it does the recordings. But once you start getting into the completeness of the data that they have that they can really gain value from, we're just at the tip of the surface, right? There's so much data that they have over, you know, based on a customer or a client that they're not even referencing right now that can gain them so much more insight. But the big part of it is just the regulation right now. Some teams I found, some teams that are probably more experienced in AI, they're willing to take that risk and try to develop an application to address a use case. But other companies that probably are less AI experienced, they tilt more to be risk-averse, and they don't want to take on this because they're fearful of having that data be exposed. Because, you know, HIPAA regulations, there's all these regulations, especially in the healthcare scenario. But at some point, someone's going to have to take on that ability of running an LLM, making sure they have the right guardrails in place and the right security for this to make some progress.

(Joel Beasley at 00:15:17) Yeah. I have noticed, my parents are in healthcare space, and I do notice that there's a lot of people that are in the middle, like, we'll see what happens. A lot of people are very scared, and there's like one or 2% that are like, we can do this. We can find the right guardrails. We can figure it out. And they lead the way. They do it. They have success, and then they share their learnings, and that makes the rest of the group more comfortable. Now, you've mentioned, you're a very bright guy. You mentioned strategy a couple times in this. And my mind wants to know what a good strategy even looks like. Is it a 100-page document with, you know, 50 people involved? Is it a one-page thing that has three points? You see all types of strategies. Is it a written document? Let's start there.

(Brian Guttadauria at 00:16:02) Yeah, it's a written document with, you know, typically get a few people around the table, you know, picking their brains, thinking about what's going on in the industry in general. But, you know, it starts from a couple different angles. And this is why, you know, I guess, from HPE's perspective or anybody in the, we believe hybrid cloud is the correct model.

# Transcript Section 2 of 3

(Brian Grettadoria at 00:16:30) Right? Because you're going to have data, you know, legacy data that's on-prem. You're going to want to, in a lot of cases, you're going to want to move the AI to the data as opposed to moving the data to the AI in the cloud. And you're going to have various different paradigms around that for sure. Some people typically get started with their AI applications in the cloud, but once they decide to run them 24/7, in some cases, it becomes cost prohibitive.

(Brian Grettadoria at 00:16:59) So they decide to run those workloads closer to the data on-prem. But the strategy in a lot of cases revolves around, first off, the data. You know, how the data is laid out across your data estate, understanding it, maybe having a data fabric that allows you to easily connect to it in a secure manner. That's typically the first step. Based on that, understanding the use cases, how it's going to be accessed with not only the applications, but the security around it.

(Brian Grettadoria at 00:17:38) But then it goes down to the models that you want to use. You know, is it going to be more large language models that you're going to leverage? And if you are, typically, industries, you know, it's best to hone in on one or two of them as opposed to letting every developer use different models because, you know, you send one prompt to three different models, you'll get 10 different responses. It's similar to the database days. When you have different versions of the database, your SQL commands would behave differently depending on the version of the database you were connected to.

(Brian Grettadoria at 00:18:18) Yeah. Same with models, right? There's inconsistency. So if you're an enterprise, if you want to hone in on either leveraging a Llama model or leveraging Gemini or leveraging Bedrock, whichever the model is you want to use, it's good to have your strategy around that and what the benefits are in the LLM space. There's also importance around if your strategy revolves around developing small language models that do specific use cases that may have better performance. So a lot of it's going to come down to what the real use cases you're trying to address, the cost is part of it, and the security—that all kind of has to be understood end to end. And the other thing that's changing in this industry around AI is it cuts across all of the traditional operational functions. Like, you'd had your SysOps, your DBOps, your network operations team, your CSO.

(Brian Grettadoria at 00:19:22) You know, you either need to bring them all together to agree on what you're doing in this strategy or maybe build out a new team that can kind of virtually work with these teams to get this work done because it is transformational.

(Joel Beasley at 00:19:38) So it must look different everywhere. Some companies might be saying this is our AI—because that's such a big term, AI strategy. They might say at company A, when we say AI strategy, we're talking about how the engineering team will adopt AI technologies. At company B, it could be about this very specific key component of their business and how they're approaching it over the next 12 months. So it can mean a whole lot of different things. Is that right?

(Brian Grettadoria at 00:20:03) Yeah. Every team has an AI expert these days, which is, and, you know, and what I find is you don't want to stifle that, right? Even within HPE, there's a lot of teams that are going off doing their, what do we call it, shadow AI functions. You know, in the initial beginnings, it's fine. But at some point, you want a holistic strategy end to end at the corporate level to really lock down how people are moving data around, how it's secured, and then associate that grander enterprise strategy to those individual BUs that may have specific use cases. So it definitely is important, especially when people start paying for different models in different locations and using them different ways. You have to kind of bring it together with some thought.

(Joel Beasley at 00:21:04) What is your AI strategy in your personal life for Brian?

(Brian Grettadoria at 00:21:08) My personal life?

(Joel Beasley at 00:21:09) Yeah. Yeah. Forget all the work stuff, all the GreenLake. Like, how are you thinking about it when you sit down at night? You're like, how do I stay up to date on this? How do I play with this tech? Like, how do you think about your strategy?

(Brian Grettadoria at 00:21:21) Well, I'm, I guess, from a, you know, being the CTO of hybrid cloud, I'm always—and, again, I've always been interested in technology from—I've been on CTO boards at various companies at, you know, EMC, Oracle, Lenovo. So I'm passionate about technology. The way I kind of stay up to date is I'm constantly reading articles on the evolution of the use of it. I'm always tinkering with it, right? So I'll open up my, you know, either Cursor or I'll open up Visual Studio Code, and I'll code up little applications. I leverage a lot of the models on Hugging Face myself. So I'll go there, start running a model, you know, depending on which, what's the API that I'm using, I could create little applications just to stay abreast of how it's working, how it's used, how the industry is changing. And then I look to see how to take the learnings I have there and push them into products that we could potentially deliver. So that's my strategy.

(Brian Grettadoria at 00:22:33) I try to stay ahead, like, when I mentioned earlier when we came out with adding the MCP support to our storage array. This is something that, you know, I read a little article, saw something we could do, put together a little POC. I have an innovation team that reports to me, and we do this type of innovation work. We kicked up a little pilot, put it together, and it became part of the product within probably within two months, right? We added it into the product. And we used AI to develop it. We used AI tools with GitHub Copilot to develop some of it. We used GitHub Copilot to create some of the unit tests around it, and it gave us the ability to come from something from an article that I read to having a team member work with our storage group to put together the pilot to now move it over to be productized in a relatively short amount of time. And this is what I can see more and more happening going forward.

(Joel Beasley at 00:23:37) How long have you been at HPE?

(Brian Grettadoria at 00:23:41) About a year and a few months. About a year and a half now. Yeah.

(Joel Beasley at 00:23:46) So they—

(Brian Grettadoria at 00:23:46) What's that?

(Joel Beasley at 00:23:46) Like, it's unusual when I hear people tell me these large organizations moving this fast. And I was like, has it always moved this fast? Is that HPE's culture? What's going on there?

(Brian Grettadoria at 00:23:56) I see a little bit of a culture question there.

(Joel Beasley at 00:23:58) I guess. Yeah.

(Brian Grettadoria at 00:24:00) So HPE itself just celebrated its 10-year anniversary just recently. They spun off from HP, spun off into HPE. Now the hybrid cloud organization itself, which I'm a member of, has only been around for about probably about three years in total now, and I've probably been here half of that time, right? And, you know, what I've tried to put in place is some of the learnings I've had over the years to try to improve execution time, improve the use of tools.

(Brian Grettadoria at 00:24:42) Like, when I first got here, there's only 20% of the team using GitHub Copilot. We're now up to about 92%, the last stats that we put together. Now that just means that the tool with the license is installed. It doesn't mean they're actually using it. But that's the next step. Like, I put in a program called RISE, the RISE program, to interview people on how they're using it, you know, have some, we call, we have little lunch seminars where we talk about the technology to promote it. I put together a little newsletter that we're promoting it just to get people to use some of these AI tools more and more within the process. But in general, the culture is very collaborative from where we are. Like I said, it's fairly new in its formation, and we're trying to drive more execution and more velocity as we go into this AI world moving forward, I guess.

(Joel Beasley at 00:25:48) The RISE program is interesting to me. So tell me just a little bit more about that. I got some questions.

(Brian Grettadoria at 00:25:55) Yep. So the focus behind that was just to drive the adoption of AI and AI tools for developers, right? You know, most developers are set in their way, especially if you're doing either microcode or things on the data plane. You have a certain way of doing things. So we put this process in place. And, again, we also work with Microsoft and the Microsoft GitHub team as well, GitHub Copilot team. You know, they had a lot of trainings, a lot of lessons in place. They came and visited with our organization and gave some hands-on training to the team that was valuable. But the real focus was finding out, you know, I would meet with the various teams to see how they were using AI, how they were adopting it, and I'd spotlight it.

(Brian Grettadoria at 00:26:46) I'd ask them, put together a little video. Took the video, put it in the newsletter. You know, they're showing how they use it to dynamically create unit tests for their product set or how they use it to do code reviews and how it saves them time. What we've been doing in the past few weeks is trying to identify true development velocity. By that, I mean, okay, if we look at Jira and we look to see how many story points were completed, you know, last year this time versus now with people using Copilot, we should see an increase in the number of story point completion in theory, and we have. We saw about a 20% improvement. Now, anecdotal, you know, we don't know how difficult those features were, whether they were meaty features or they were just bug fixes. That we have to dig into a little bit deeper. But on the surface, we have been seeing an improvement in the amount of work that's getting done using AI.

(Joel Beasley at 00:27:57) That is super cool. You know, you're the second person I've talked to that's been doing things like that. Like, everyone has their own version of it. But that RISE program, the idea that you're taking the best use cases of people organically happening within your company and then highlighting it to other people. That really hits home for me because, you know, I'm a software engineer, over 20 years, and before LLMs came out, and when LLMs first came out, I started using them and they obviously rapidly improved. So my first initial touch with them was like, okay, it's all right. And then they got a little bit better, and then one of my friends sent me a video on YouTube of an engineer doing something and I said, oh, I'm using this tool wrong.

(Brian Grettadoria at 00:28:42) Yeah.

(Joel Beasley at 00:28:42) Like, it's not a one-shot app tool. It's not, it's like, you have to use it in this—you watch it and you just understand how to use it. And I'm like, oh, okay. So within five minutes, I went from like 1% competency with LLMs to like 70% competency just because I saw another person that had the same experience level that I had use it. And so that's so valuable.

(Brian Grettadoria at 00:29:06) Yep. That was the exact model. I figured, why listen to me give you something in an email? I'd rather you see your peers doing it because they're in the trenches, right? They've been doing some amazing work with it, and they're well ahead of the curve with using it, more than even myself.

(Joel Beasley at 00:29:24) That's brilliant. Well, I wanted to chat with you too about the real-world impact and customer stories. What is the coolest thing that you've seen a customer do with GreenLake Intelligence?

(Brian Grettadoria at 00:29:35) Two things are really cool. One is that migration use case that I was referencing in general, being able to do both the planning stage and the day-one migration activity automatically. The next piece that's very, very interesting with us and we're seeing customers do this right now—I don't know if I can mention the customer or not, but let's say a large home product retailer. They use it for their day-two analysis. Basically, going through their observability data to do root cause analysis of problems that they're seeing in their environment.

(Brian Grettadoria at 00:30:15) Think of it as AIOps 2.0, right? Leveraging an agentic interface to go through that data real time, you know, leveraging GreenLake Intelligence and the GreenLake platform where we're collecting this information historically over time to give you a view of what this failure really is and how to root cause solve it, right? These are some of the cases we're seeing today. The next things from that, what we're working on is integrating this into third-party tools. So by that is integrating it into ServiceNow. Now all the operations we do perform, for instance, we can automatically create change tickets for it. So you can see, you know, in your CMDB that this actually did get updated, or it can send a notification to said teams so they can double back on it and update their documentation, for instance. So these are some of the use cases we're seeing really resonate with our customers today.

(Joel Beasley at 00:31:13) And then what's your job like? Do you spend a lot of time with customers talking about this, like, how they're using it and all that?

(Brian Grettadoria at 00:31:19) I actually do. So, you know, like I mentioned last week—not last week, week before last—when I was in London, we were visiting with customers. They were in the process of defining their AI strategy. You know, part of it was like, what are some of the use cases we're taking from the business unit? How can we apply it in our organization? They were building out some AI infrastructure on-prem. They actually used a competitor's device, and they were not very happy with it. It was interesting to hear some of their—yeah, it was interesting to hear their feedback.

(Brian Grettadoria at 00:31:55) They started early with a competitor's product, which was more of a reference architecture. They spent more of their time actually trying to get the device up and going than time spent on the use case itself. So I mentioned our product, Private Cloud AI, what it's designed for—it's time to value, get you up and going with the product relatively easy. We already have the NVIDIA AI Enterprise built into the stack. So you turn it on, and you can start using the blueprints to create your ML workflow. And they were really interested in that. It was like, oh, like, we should have had this a few months ago. But then we went into, you know, I spoke to them about what we're doing in GreenLake Intelligence and how it can address your grander operational issues. And they mentioned they were in the process of building out three new data centers, and this really resonated with them because they're looking for a way to uplift the way they operate these data centers. They can't be doing it the ways they were doing it in the past. They need a new model.

(Joel Beasley at 00:33:00) Be like, I'm just going to ship you some equipment. Yeah. And we can just pay the invoice. Yeah. We'll get it taken care of.

(Brian Grettadoria at 00:33:07) But a lot of this is, you know, it's still—AI is very new. It's still evolving. We're also looking to work with, you know, customers on this joint innovation. There are probably use cases that we're not even aware of yet that we'd love to explore with our customers as we build this thing out. So I guess that would be a call to action here.

(Joel Beasley at 00:33:28) Are your customers typically finding the use case and then approaching you for the technology to help them execute? Or are they coming to you? Do you have any consultative arm where they're like, hey, we know we need to be involved in this, but we don't know how it's going to apply to our business, or we have five ideas of things we could do, which one should we tackle? Do you guys do that side of things, or you just provide the technology?

(Brian Gruttadauria at 00:33:51) So we do both of those. We have a series of customer advisory boards where we gather this input of how they're either using our product or just how their challenges in general and how we address it. They do it formally at the executive level with some of our large customers. And then we have another organization that focuses on more smaller, they call it CTAB events, where they will have a roundtable, meet with like a half dozen customers. They'll whiteboard their challenges, and we try to come back to see how we can add this into the roadmap.

(Brian Gruttadauria at 00:34:37) One of the reasons that's important, and I'm glad you brought that up, is in addition to having the CTO council that I formed up, we've also built out this product management council where we get all the product managers from all of the organizations across Hybrid Cloud together to pontificate about our roadmaps. But it's more than that. It's like going through the roadmaps, identifying features, and then we bring in some of these challenges to see how best to address it and to remove duplication. Right? Because we have several different, whether it's OpsRamp, Morpheus, et cetera, we look for opportunities to drive these challenges that we've identified from the customers directly into the roadmap.

(Brian Gruttadauria at 00:35:26) Right? And part of that is done where we can get it into a roadmap on a regular cadence. In other cases, they may not have the bandwidth, and my innovation team will look to try to create an early MVP to shake out that use case, see if it resonates so we could hand it off to the engineering team. So that's the motion that we've put in place. It's a similar motion that I kind of had at different companies in the past, and I've seen instances of it where it worked, instances of it where it didn't work.

(Brian Gruttadauria at 00:36:04) In this case, it's been pretty good. We've probably over the past, I guess, probably eight months, we've had about six or so innovations that have come from my team directly go into the roadmap, which is a pretty good percentage. Right?

(Joel Beasley at 00:36:24) That's pretty cool. That's exciting, man.

(Brian Gruttadauria at 00:36:26) It is pretty cool. I have a, let me, I could send it to you another time, but I have a tech radar where we have about a dozen or so innovation activities spanning, you know, some of them just simple investigation. Other ones are putting together a POC, and they're focused in different areas. And these areas are focused to try to address challenges in a lot in some cases, and other ones are innovating with our partners on things that are on their future roadmap. So for instance, working with NVIDIA on something that's on their roadmap.

(Brian Gruttadauria at 00:37:07) Customer probably won't see it for another year or so, but we're starting to kick the tires and look to see how we can productize it.

(Joel Beasley at 00:37:14) Is this tech radar, this is a management thing, or how does this work?

(Brian Gruttadauria at 00:37:17) This is just something that helps me track and visualize all of the POCs and innovation activities that we have going on.

(Joel Beasley at 00:37:27) Document you keep personally to be a good executive.

(Brian Gruttadauria at 00:37:31) Yep.

(Joel Beasley at 00:37:32) And you call it tech radar? That sounds really cool.

(Brian Gruttadauria at 00:37:36) I think it's something similar to, you know, Gartner has something similar to this. They put together these tech radars where they're looking at the technologies and how they, you know, you're looking at things that are three years out, two years out, one year out. So as they progress from, you know, things that we believe are a little bit further out and they get closer to being productized, they'll move in on that radar.

(Joel Beasley at 00:38:03) That is cool. Did you have any part of the Juniper acquisition, or did you just, it just happened at the company?

(Brian Gruttadauria at 00:38:09) It was already in flight before I joined.

(Joel Beasley at 00:38:14) So you closed the deal. You made it happen. Brian made it happen.

(Brian Gruttadauria at 00:38:17) I'm not gonna say that. But we are working very closely with that team. You know, had a really good relationship with Raj, the previous CTO of Juniper. You know, we still stay in touch now. He's a great guy.

(Brian Gruttadauria at 00:38:36) And we discussed several different technologies, one of which their Contrail CN2 stack, which is their software-defined networking stack. We were, you know, first in working with them to find an innovation, working with their team. They have a really solid engineering and innovation team there to take that CN2 stack, and we're building it right into our VME, basically VM Essentials stack. So very soon, I'm not gonna share any dates, but we are rapidly looking to integrate that software-defined networking stack into our VM Essentials.

(Brian Gruttadauria at 00:39:15) One thing that I'm trying to do here is, like, people look at HPE. They look at it as like, oh, okay. It's a hardware company. Right? I would argue our organization and what we've built, at least since I've been here, it's a software company. All of the people that we have, a large lot of them are software engineers at their core. So we're right in there.

(Joel Beasley at 00:39:52) Right. So, yeah. Well, I knew I liked you when I read that you were using AI to analyze real estate investments and cash flow. Because I'm like, that's a nerd. That's what we're doing.

(Joel Beasley at 00:40:01) Because I grew up in real estate. My parents were in real estate, and just that weekend hacking a project for cash flow analysis with AI. I'm like, that sounds completely doable. That's a Saturday afternoon thing now with as advanced as technologies have gotten with vibe coding and all of that.

(Brian Gruttadauria at 00:40:16) Yep.

(Joel Beasley at 00:40:17) So I love it. I like to use it.

(Brian Gruttadauria at 00:40:19) That's what I've been doing for the past year or so. But it's my Sunday morning cup of coffee. Sit down with the MLS, and I analyze them to see.

(Joel Beasley at 00:40:31) That's right. Yeah. You got a friend. You got that RETS API.

(Joel Beasley at 00:40:38) Oh, okay. Yeah. And just for people that didn't know, I didn't get RETS wrong. RETS is actually the Real Estate Transaction Standard. Yeah.

(Joel Beasley at 00:40:44) Yeah. It's important to me that some people heard that. They're like, he meant REST. Right? And it's like, no.

(Joel Beasley at 00:40:50) I meant RETS.

(Brian Gruttadauria at 00:40:53) Only people who dabble in real estate know that.

(Joel Beasley at 00:40:55) Yes. Yes. Yes. You should have seen my face the first time, Brian. I was at a conference fifteen years ago, a real estate conference, and they were talking about the RETS.

(Joel Beasley at 00:41:05) And I had experience, you know, integrating it already. And I thought it was so funny because I learned the acronym at the conference, and I was like, standard. It's a standard, but there's 700 versions.

(Brian Gruttadauria at 00:41:15) Yeah.

(Joel Beasley at 00:41:16) It's like completely, it's like, what am I supposed to do with this? You might as well not even have the standard. You know? Well, let's wrap up on some leadership insights. You didn't get to run, be the CTO of HPE without a lot of experience growing, managing teams, and interacting with people.

(Joel Beasley at 00:41:30) So I'm just gonna ask you my favorite leadership advice question, and we can kick off from there if that's okay with you. So I'm gonna ask you for a piece of advice, but there's gonna be some constraints. And the constraints are, here we go. Somebody gave you this advice.

(Joel Beasley at 00:41:45) You implemented it. It worked extremely well, and you've kept it in your toolbox for a long time. What advice was that?

(Brian Gruttadauria at 00:41:55) Probably the best advice I got, this is years ago, my days at EMC. Well, you know, one of my mentors. I still meet with them at least once or twice a year. His name was Joel Schwartz, and I worked for him at EMC. I had an opportunity to take two paths at one point. It could've went the safe path and, you know, manage a small team or taking a path where it kind of was much riskier, much more innovative.

(Brian Gruttadauria at 00:42:34) But it was high risk. So his suggestion to me was, oh, yeah, especially in the technology industry, don't be afraid to take a risk in your career. It'll lead to better opportunities. And that's the kind of advice I would give to anyone in the tech industry. It's always easier to continue doing what you're doing right now, but try to push yourself and take that challenge.

(Brian Gruttadauria at 00:43:02) Like I mentioned in the past, I always kind of incubated different ideas that led to the formation of teams, that led to product or the acquisition of a team, or kind of led to a new joint venture. Something I always enjoy doing, and it was largely based on that advice I got a long time ago. So that's the advice I would say. Don't be afraid to take a risk.

(Joel Beasley at 00:43:28) And then what's something that when you're hiring, bringing executives onto your team, there's a million things you could look at. Right? But what's the one or two things that you really, really hone in on when you're looking to bring a new member on? Like, what traits would they have?

(Brian Gruttadauria at 00:43:48) That they'd be inquisitive and be willing to implement change. Right? Our industry is all about change, change management, being able to deliver and drive a change, and especially now. I've hired just recently within the past few months, I've hired about three people. And I always look at them as, you know, have you on your spare time been tinkering with, you know, if I can go to Hugging Face and download a model, I would expect someone who works for me to do something similar as well. But to be able to drive and implement that change and not be adverse to trying something new.

(Brian Gruttadauria at 00:44:30) Right? I'd rather someone have the ability to fail fast, be open. That's another key thing. Be open with some of the challenges they run across and not be afraid to come to me with things that are not working, and we can discuss how we can fix it together. But I would say that the biggest trait would be open to implement change and be a change driver.

(Joel Beasley at 00:44:57) I love that. And, you know, you brought up a good point about the being open to discussing things that aren't working and bring them to you. For me, when as growing and hiring and building teams, that was the biggest surprise to me. Because I don't know what it is about me or how I was raised or genetically whatever you want to call it, but I have always been like if there is a problem, we got to just run it up the ladder because we are trying to achieve the solution. The goal is the solution.

(Joel Beasley at 00:45:20) The goal isn't me looking awesome. The goal isn't me trying not to make a mistake. The goal is to hit the target and we need to bring in everybody and swarm this problem. And then, I found out really quickly that's not the default nature for most people. Most people want to hide.

(Joel Beasley at 00:45:34) They want to retract and solve it and then come back and, like, it's, you know? And it's a delicate balance between that. Right? But—

(Brian Gruttadauria at 00:45:43) It's a culture you need to build in because I found in other companies I've been in the past, not HPE, but other companies I've been in the past, it was easier to say no and easier to just kind of stay in your lane than stick your neck out. Right? So at least here, I try to encourage people, you know, don't be afraid to bring it up. You're not gonna be punished for it. Right? We can discuss it, go through it.

(Brian Gruttadauria at 00:46:09) Certainly, there are some situations that deem some action. But in a lot of cases, it's, you know, like you said, we're trying to achieve a better goal together.

(Joel Beasley at 00:46:22) If you're a high quality person, you really care, you're constantly improving, then raising the problem is a good thing. And then also then, Brian, worst case scenario, you raise the problem, you get your hand slapped, and then you just learn how to read the room better. Yeah. Which is its own skill.

(Brian Gruttadauria at 00:46:40) Yeah. It is too. You are right. Send it via text next time.

(Joel Beasley at 00:46:45) Yeah. Yeah. Choose the words wisely, you know. Cool. Brian, we did it man.

(Joel Beasley at 00:46:49) We made a podcast. How do you feel?

(Brian Gruttadauria at 00:46:51) We did. I feel great.

(Joel Beasley at 00:46:53) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you would like to hear discussed on the podcast, either add me on LinkedIn or send me an email [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.