Matt Ausman, CIO of Zebra Technologies, on AI Agents and the Future of IT

0:00How do we ensure we're even building the right things? Just because we can build it doesn't mean we should build it.

0:07Peter High: Welcome to Technovation. I'm Peter High, president of Metastrategy. My guest today is Matt Ausman. Matt is the chief information officer of Zebra Technologies, a company that helps organizations digitize and automate operations through mobile computing, scanning, printing, and software solutions, serving industries that include retail, manufacturing, transportation, and healthcare. Zebra Technologies earned approximately$5.4 billion in revenue in 2025. Matt leads the company's enterprise technology organization, partnering with business leaders to modernize operations, strengthen its data foundation, and integrate AI into everyday work.

0:41Peter High: He joined Zebra Technologies just over five years ago, and he became CIO a little more than four and a half years ago. He spent most of two decades at General Electric, ending his time there as a field services CIO for GE Healthcare. In our conversation today, we'll discuss how Matt is preparing for a workforce that combines employees, service providers, and digital workers. We'll discuss how AI could reshape software development and organization structures, how he approaches platform decisions, and how Zebra's Customer Zero Program informs technology adoption, among other topics we'll cover together.

1:12Peter High: Matt, welcome to Technovation. It's great to speak with you today. Yeah. Thanks, Peter. Glad to be here. I'm glad you are here. And now for a word from our partner, Blitzy. Want to shift software five times faster? Meet Blitzy, the autonomous software development platform built for enterprise scale initiatives. Your code base has outgrown any one team. Tens of millions of lines, decades of technical debt, more than any other AI tool can hold in its context window. Blitzy starts by understanding. Thousands of specialized agents map your architecture, dependencies, and business logic across every language into a knowledge graph.

1:47Peter High: That graph also services hidden CVEs and zero days before an alert ever reaches your desk. Then you describe what you need in plain English and approve the agent action plan Blitzy builds. Grounded in the knowledge graph, Blitzy executes entire engineering epochs over weeks of continuous compute, and every line is tested end to end before it reaches your team. Modernize legacy systems, close critical vulnerabilities, ship new features five times faster. Learn more at blitzy.com slash technovation. That's B-L-I-T-Z-Y AI.com slash technovation. And now on to the interview. Let's begin with Zebra Technologies.

2:27Peter High: I provided a thumbnail sketch to the business, but for those who may be less familiar with it, provide your own background into the business that you're a part of. When I think of Zebra, I wish that I knew about Zebra before I came to Zebra. It's a hidden gem. And what we really look to do is provide solutions for frontline workers. So we want to be the AI leader for the frontline. And what that really means is how do we put the right hardware, the right software, the right solutions for different workers? So in my time at GE, I led field service teams around the world and we're monkeying with devices on our arms and hanging them.

3:06How do you put a device in their hand that has the right fit, works around the world in all the different temperatures. So we create the solutions for those regardless of the industry. So whether it's healthcare, manufacturing, transportation, logistics, retail, putting a handheld in their hand that can get them access to the information they need, tracking that package from where it's manufactured all the way to your doorstep. Or in the hospital, when I spend time at GE Healthcare thinking about what it took to fix an MRI machine or a CT scanner, but what does that nurse need to be able to deliver care to that patient.

3:43So how do we put the right technology in their hands, access to the right data, whether it's using cameras or vision, RFID, many of the other different technologies that surround those frontline workers?

3:56Peter High: Fascinating. Very interesting overview. I appreciate that. Talk a bit about your role as chief information officer. I mentioned at the outset, you've had the role for a bit more than four and a half years. What does your role entail. Yeah. I continue to wake up every day energized around my role. I grew up with a lot of time in the enterprise solution space. So think of all the applications that our team, that our functional teams need to use to get their job done, but then infrastructure, all the cybersecurity needs of the company. We cover all of that. Data warehouse. What do you need to provide data across the entire company?

4:29How do we move data inside and outside of the organization, then there's your normal help desk, making sure that we all have working laptops, conference rooms, collaboration tools, and more recently, spearheading AI within the organization. So it's the buzzword that everybody talks about. And I'm very lucky and happy that our leadership team has trusted me to help drive the organization forward in how we use and adopt AI internally.

4:56Peter High: I look forward to double-clicking on that thoroughly. Before we do, though, talk a bit about the way in which your organization structured, Matt. Talk a bit about the team that you lead. Sure. So I can say I'd split the organization across what I call IT enterprise solutions leaders. So these are leaders that interface directly with their functional counterparts. So HR, finance, marketing, sales, finance, they all lead different solutions, whether it's the ERP, the CRM, legal tools. Those leaders are business functional facing, driving productivity and value across all the employees in Zebra. Our product and solutions team, you name it, every one of them needs tools to do their job better so that we can provide for our customers.

5:45Then there's the other half of the organization that I've got split in a couple of different areas. We've got enterprise architecture, data and integration. So what do our system architectures look like? Where is our data structured? We have a centralized data warehouse, but there's a lot of data in the company. How do we move that data around? Then we've got our platform teams. This is your infrastructure, our data centers, our cloud platforms. How do we put the right guardrails in? We've got an innovation team. They support us across the organization. And then running overall IT operations. So financials, program management, et cetera.

6:21So I look at it as there's the functional facing teams, and then there's the internal, how do we run IT for ourselves? And what do we need to do to keep the company going?

6:31Peter High: And I know when you joined the firm, Matt, it was a heavily outsourced team. And that's something that you've changed since coming into your role. Talk a bit about rebalancing that, the reasons for doing so and where you are in that rebalancing? Yeah, we've been on a really great transformation over the past four and a half years. A couple of things that we did, we transitioned to the model with the IT enterprise solutions leaders, really getting away from just the BRM and delivery lead. I want to create leaders that are facing and own delivery. But as we got into that, we realized that we were outsourced almost 80%.

7:07So leveraging managed service partners, they were there for us, but we really felt the need to do more ourselves. So over the past four and a half years, we're closer to 50-50 model now, which I feel a lot better about. And with how AI continues to take some of the work that is done at different levels, it's giving us that expertise to be really into the weeds and details. When I think of our organization, our structure for me is really important. So when we all think of a company, we think of the pyramid, right? That's the structure of the organization. And when I took over, the IT employee was really a diamond.

7:48So what does that look like? The two corners were outsourced. So wide in the middle with senior managers going down to very few entry-level employees. And what we sought out to do was transform that. And I think we've changed our shape to be a little bit more like a house. So take that diamond and spread it out. We invested heavily in our entry-level program. So we've got an entry level leadership program in the U.S. for college grads. And we've ramped up our hiring in our India Development Center that we have. So we've been able to grow talent, bring in source some of the capabilities and functions.

8:26One example is within Salesforce. We're really excited that we've got a team that can do the majority of the work we need internal. It's a core platform for us. We've got the right skills. We've got the right architecture. and we can deliver so much more faster, especially with AI coming into the mix. So we've gotten it to where we feel good about it being a house. But I do think about as we transition to an AI first world, what does it look like when digital workers come in? Will they just replace the corners and our in-source, out-source ratio changes? We start saying, what's our digital to human worker ratio?

9:04Or does the shape change again? Does it become almost like a little funnel where it's skinny and narrow of what we have at the lower levels up? Is it a regular funnel? Like really thinking about what the structure of our internal organization paired with digital workers is really fascinating. It's one my team and I are spending a lot of time on trying to figure out what that shape of our internal versus external team should look like.

9:33Peter High: Yeah, very interesting. I look forward to hearing how that continues to evolve and how they're incorporated in. You mentioned earlier, Matt, that you wish that when you were GE, you knew about Zebra to become a customer of it. Well, you are now a customer of it through the Zebra for Zebra program, sort of a customer zero program, as others may know it. Talk a bit about Zebra for Zebra and the role that you and your team play in the course of that and helping, in some cases, shape the products that the organization is developing. Yeah, it's something I'm really proud of in the work that our teams have done to collaborate with our product teams and with our internal organization.

10:11So we run a couple of distribution centers and it's one of the best test cases where we bring our products in. So we were one of the first to launch our Zebra Companion. What is that app that helps your new workers that join our distribution centers? How did they get the help that they need? How do I put something away? Where does it go? multi-language. It's one of the things we're able to help actually contribute to the product team. For us, we really wanted multi-language and we're able to develop some code and actually give it back to our product team to absorb. But also there are requirements.

10:47We're working on one where using cameras to scan a bill of lading or what's coming in to be able to do that, take a picture of a palette and be able to auto-receive. Really great use case using our cameras and some of our machine vision, but the initial product worked with just a one-page scan of that document. We don't just have one-page scans. We have multi-page. And so pushing them a little bit further to say, you're going to need to do this with your customers. It's great for us in proof of concept. I can go and take a picture of one and go through it, but 20 times I've got to do that. I want to just scan 20 pages and then take a bunch of pictures.

11:32That's the real transformation of the use case, not scan a picture, look for the palette and the humans still doing a lot of work. So pushing our product teams to really understand some of the use cases and what really happens is great. And where I talk to our teams in IT and the product teams, theoretically, we're not going to fire the business team and hopefully they don't look at us as too needy. it's that symbiotic relationship where we can work and help them give them a shape and a direction. If the capability doesn't come soon, we can sometimes build it ourselves and give it to them. That's a great way that we can work and partner together.

12:11So we've really pushed the limits in RFID the last couple of years of being able to show how we can fund it and grow because that's helpful for our sales teams as well when they're going out talking to customers. What's the payback model? What's the ROI? How do we use it? How is it valuable for us? So really looking at that 360 view of the cost, the benefits, the requirements, the capabilities, and being able to really drink our own champagne.

12:41Peter High: Yeah, it's interesting. Last time you and I had a chance to catch up, Matt, you had a comparison of this moment in AI and the transition from waterfall to agile development. It was an interesting analogy you had, and I wonder if you'd share it with the audience here. Yeah, I think about 20 years ago, and many of us were running Waterfall, and it was the greatest. It was what we all did. It was best in class. Agile came out. When I was in college, I learned a little bit of extreme programming, and Zebra made the shift to Agile about five years ago. Earlier in my career, I'd done it, but I feel like in some companies, it's only been around 10 years.

13:17But I think about 10 years from now, what is going to be that operating model? I wish I could come up with a cool name for it. Write that book to say, what is the scrum team of the future going to look like? Are we going to have scrum teams, right? We used to have waterfall with business analysts requirements and how we went through it. We're now at a model with a scrum team that you've got your technical and your functional partners aligned. But what happens in 10 years when is it just a functional product owner and a technical product owner? and they can do seemingly limitless work. And what ideas we dream up can become reality in short periods of time.

14:02So I don't know what that change management model looks like. The speed at which you can go will probably eclipse the change management your organizations can take in. How do we ensure that we're designing safely? How do we ensure we're even building the right things? Just because we can build it doesn't mean we should build it. for our functional partners and with them. So it's going to be interesting when we look back 10 years from now and say, what's the new methodology we're putting in? Who's the group that kind of figured it out, wrote the book that we all, that playbook that we all run from?

14:40Peter High: Exactly. Well, you consolidate analytics and data and the data warehouse onto a single platform. And I wonder what changes in data ownership or information flows and team responsibilities made that successful. And I'm also curious of what opportunities AI now has as a result of your having made these changes. It seems like you set a new foundation that AI then can build upon, in essence. Is that fair? Yeah, it's been a fun but challenging transformation in the data space. So over the last four or five years, we've consolidated our data science platform where we've truly got many different types of data that our data scientists are out playing and thinking up new great things.

15:27In the same space and connected, we've got our enterprise data warehouse. So you've got that data coming in. We've got linkages where how do we pair that with our telemetry data? What's going on in the organization? So it's been really fantastic to see it all in one platform. But doing that has really required us to think about how we manage data. So in Zebra, we've got a federated model for our data. So I would say our IT organization runs and owns the platform, but supply chain's got data. They need to manage it. They need to understand who's building reports. Why are we building the 2000 finance report that is just like the 1999th report, right?

16:10That governance, how do you make sure that you're not recreating things, building layers upon layers of abstraction. So each of our functional teams has realigned. So we've got data owners for each of those functions that provide a level of governance, who's accessing those models. And in some respects, what it's forcing is the cross collaboration of functions. It's easy in your own function to say, it's finance, I'm going to build a report. But what happens when sales needs that data? Sometimes they like to just copy it, create it, build new. But how do we stay on those same data sets so we've got data integrity?

16:50Because you mentioned AI, that's where it really starts to be powerful. But running AI on Excel spreadsheets worth of data that are on a laptop or unvalidated data is going to give you bad results, just like we may have gotten bad results five years ago. When somebody goes and morphs or comes up with a new definition of what does sales data mean? Is it daily? What is it? Oh, I'm going to exclude these five things. I'm going to include these. When AI goes and tries to figure it out, well, it's no surprise when we ran some initial AI tests and AI told us that one of our business segments had$9 billion of revenue, which you mentioned the whole company only has five points.

17:34How you get that data right is so important with AI because humans aren't always checking. And where I get really nervous is the small blips. The big ones are easy to catch. We're never going to miss that a BU segment had$9 billion of revenue. I am never worried about catching that. We will always catch that one. It's when it's off by a million dollars. It's off by$2 million. Those are the ones that may slip through and you've really got to govern your data well. And that's the partnership that IT has got to give the right platform, the right tools, the right structure. And then our business functional counterparts really own the data that comes into the system.

18:17And how do we ensure that it's not morphed and modified in ways that we all don't know and understand? end.

18:24Peter High: Speaking of those business functional counterparts, you've emphasized to me that functional leadership is essential to AI success. You're really excited about what's brought to life through the collaboration between you and those functional leaders. Talk a bit about some of the exciting developments happening across Zebra in some of those functions. Yeah, I believe that for us to add value to the company, it takes strong technology, deep business process understanding and people. It's a three-legged stool, people, process, technology. You probably hear that all the time. And it's not one or the other.

19:01It's not IT alone by themselves, our functional teams. It's really a collaborative partnership. So what I've started to see is our functional teams are able to do more and more with technology. You look at some of the newer coding assistants that are coming out and people are building things. And it's interesting because it's even pushing the limits within some of our functional teams. So in the past, we all can talk about shadow IT. That's happened in CIOs and IT leaders have seen that for years and years where you've got teams that start building things or access databases and VBScripts running all over.

19:38And a lot of times that got centralized. So you had techno functional teams. And I looked at that and said, well, the reason that happens is because IT cannot adapt quickly. We don't understand enough of the context of the business. We're not able to meet them with what they need. So now, though, you take some of these tools, any salesperson can go and build an agent that's helping them do pricing on their deals. But is that what we want? You've got 50 people doing that. And I look at some of the functional leadership looking and how do you control that, right? Anybody can, with some training, develop their own tools.

20:18And it's a little scary. So we spent a lot of time trying to figure out what is our funnel? How do we work from a leadership team down to ensure we've got to bring the ideas in? We've got to understand what the sales reps, what the pain they're feeling, our people on our product development teams, what's slowing them down. So we've got to hear that. That's the reason that people want to build things. If we gave them the tools that they need, they wouldn't go off and be customizing and building necessarily. So we're really trying to ensure that we're taking the learnings from what people are doing individually.

20:55And we're actually doing that with agents today. So we've just launched an agent builder within Zebra. And we've got 4 ,600 agents that have been built in a couple months. And I'm okay with that. I want individuals to build a bunch of agents. I want them to learn the tools. I want them to upskill. But what we're actually now digging in is you've got a team that all built almost the same agent, right? The title is the same. This group built it. How do we go and figure out the one that really should be used across all of them? But we're taking it in steps because one of the things that AI is continuing to reinforce is the need for governance.

21:34Then how do we manage and maintain these agents? We've had these constructs over time. An Excel spreadsheet gets built by someone and it gets shared by five people and it's got formulas and it's got functions in it and maybe it doesn't get updated and it eventually breaks the accountabilities on the human. But as we continue in AI, what if I build an agent and I give it to 10 people and it gives the wrong output? Is that on me because I built it? Is it on the 10 people that are using it? There's drift in AI. There's no two plus two equals four in a very prescriptive function. It's thinking like a human.

22:13It learns based on our prompts. We can retrain it. So how are we doing that? And so we're taking some baby steps. Today, we don't allow agent sharing. You go and create an agent. You use it. You're accountable to the results. We will start to scale where I gave that example. How do we go into the different functional teams and learn what do they really want? What do they need? and let's put that governance model. Who's going to update and maintain this agent for the group? That's a new role that's likely going to come out in the future. We think about what has happened over the last 150, 200 years with technology.

22:52We continue to create new technology and say all the roles will be gone, but they don't. They shift and change. People are working on fixing robots, right? They may not be working, touching every warehouse shelf. A robot may do some of it. We've got humans programming them. We've got them fixing them. We've got them thinking of new use cases. So what are those new roles that get created over the next few years? Back to your question on what's that new paradigm? It's not agile, it's something else. Today we have scrum masters and product owners and developers. In that new model, what will we have?

23:30what will those new roles of the future be?

23:33Peter High: It sounds like you're actively brainstorming those very things with your team and starting to play out the scenarios, at least, that might determine which direction the organization goes in. Yep, definitely. Yeah, that's interesting. I wanted to ask you also, Matt, how have you thought about assembling an ecosystem of tools around you? There are a variety of things that are homegrown, But I know you also have, as anyone would at an organization the scale of yours, built out an AI ecosystem of sorts as well. What are some of the principles you've used to determine where to invest and where not to?

24:10Yeah. So at the Zebra level, we really work very hard to be single system. So as an example, we have one ERP, which is great. We have one CRM. We have one data warehouse. I'll say on one hand, there's probably a subset of, we call them anchor platforms. So we build around those. And with AI, we've got one as well. So we have an anchor platform that every single person in the company has access to. We want to ensure that there's broad-based uptake and usage. So we've got anchor platforms. And then what I say is everything floating around it is buoys. My goal is not to go and custom build or replace those anchors.

24:51We're going to buy those anchors. We're going to leverage those enterprise systems, the millions of dollars of R &D that go into your core CRM, ERP, HCM, those systems in the enterprise. But then you get a lot of buoys hanging around and we're starting to have discussions on do you build or do you buy those? And it's an interesting dynamic in today's world of what do you build and buy? The cost financials are starting to change in today's world, but it's not free. We sometimes go and say, oh, I can build this app. It's going to take us two months and we're going to build it. And great, we can go and save licensing costs from the enterprise.

25:34But guess what? You're only as good as you are today. Who's the product manager that's going to maintain it tomorrow? We've got cybersecurity needs that we've got to keep up. We've got to continue to support and grow the platform. How does it fit in with the ecosystem? So there are some hidden costs out there, but we've definitely found areas that our teams have been able to go in and show that we can go build something and replace some of those off-the-shelf, I'll call them kind of pocket apps, or it's a point solution that we may have been working to not necessarily customize, but configure and make it work.

26:13So in that vein, we're going to go and figure out what's the right way that we structure the organization from a build versus buy and what our architecture looks like.

26:24Peter High: Very interesting. I wanted to also ask you, Matt, before I let you go, what have you recently read or listened to that you'd recommend to others? Or likewise, what do you consume in order to sort of fill your mind with what's possible given all of the appropriate brainstorming you're doing with your team about what the future might hold given some of the changes we've described and discussed today? Yeah. I don't know if I've got any one good piece of material. I wish I had a podcast. I just got into podcasts a little bit more. I do love Notebook LM where I go out to the internet. I consume a whole lot of information.

27:03It's creating podcasts for me to listen to. I'm an avid reader of many different sources. And I'll be honest, I love talking to peers. One of the things that I really appreciate when we've met earlier this year is networking. I learn through my peers. I love having those discussions. But one of the things, I read it a few years ago, and we had a leadership training course at Zebra that I had the chance to speak at, but I don't even remember the title of it. It's about followership. We spend so much time talking about leadership, but we sometimes forget what followership means, right? I'm not always leading.

27:46When do I step back? How do I empower others? I think in this AI race, I'm having to learn with our functional partners, where do you have the domain expertise? Where do I want you to lead? How can I be the best follower for you? I've got experts. We've got people that are AI natives coming in straight out of college that will run circles around me and what I understand about AI and how it works. Being able to step back and follow and let them lead is something that I've taken with me over the last 10, 15 years is what does followership mean and how do you use that to actually grow your leadership?

28:22Peter High: That's fantastic insights all. Well, Matt, thank you so much for a terrific conversation. Phenomenal to learn from you and the sorts of things across your tenure at Zebra that you're implementing there to help the organization seize value in the future. It's been a great conversation. Thank you. Awesome. You're very welcome. Thank you, Peter. Thanks for listening to Technovation. If you enjoyed this conversation, subscribe and leave a rating. It helps more leaders find our podcast. Thank you. And I look forward to seeing you next time.

28:55Thank you.