As enterprises move beyond AI experimentation, the challenge is turning ambitious strategies into secure, scalable operations. What does it take to manage autonomous agents, maintain control over enterprise data and enforce governance across increasingly complex hybrid environments?
Bruno Aziza, group vice president of software, data, AI, automation and security at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating Model 2026 interview series to discuss how organizations are putting AI to work while addressing the operational demands that come with it.
The conversation covers the shift toward continuous AI governance, the growing importance of machine identity and the role of containerization in deploying AI workloads consistently across hybrid environments. Aziza also shares insights into orchestrating data, technology and operations while balancing regulatory requirements, along with lessons emerging from IBM TechXchange and real-world customer deployments. As AI agents multiply and models advance, the discussion highlights what enterprises must prioritize to turn AI investments into measurable business value.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM's AI Operating Model Series. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for IBM's AI Operating Model Series
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for IBM's AI Operating Model Series.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
IBM's AI Operating Model Series. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to IBM's AI Operating Model Series
Please sign in with LinkedIn to continue to IBM's AI Operating Model Series. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Bruno Aziza, IBM
As enterprises move beyond AI experimentation, the challenge is turning ambitious strategies into secure, scalable operations. What does it take to manage autonomous agents, maintain control over enterprise data and enforce governance across increasingly complex hybrid environments?
Bruno Aziza, group vice president of software, data, AI, automation and security at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating Model 2026 interview series to discuss how organizations are putting AI to work while addressing the operational demands that come with it.
The conversation covers the shift toward continuous AI governance, the growing importance of machine identity and the role of containerization in deploying AI workloads consistently across hybrid environments. Aziza also shares insights into orchestrating data, technology and operations while balancing regulatory requirements, along with lessons emerging from IBM TechXchange and real-world customer deployments. As AI agents multiply and models advance, the discussion highlights what enterprises must prioritize to turn AI investments into measurable business value.
Group VP, Software (Data, AI, Automation & Security)IBM
As enterprises move beyond AI experimentation, the challenge is turning ambitious strategies into secure, scalable operations. What does it take to manage autonomous agents, maintain control over enterprise data and enforce governance across increasingly complex hybrid environments?
Bruno Aziza, group vice president of software, data, AI, automation and security at IBM, joins theCUBE Research analysts John Furrier and Dave Vellante for IBM's AI Operating Model 2026 interview series to discuss how organizations are putting AI to work while addressing th...Read more
exploreKeep Exploring
What does "AI-native enterprise" mean, and how does it differ from being "AI-ready"?add
How are leaders thinking about sovereignty — do they see it only as data location or as a multi-dimensional issue (data, technology, operations, regulation, and finance), and what's your perspective on that?add
How should governance be approached for large-scale agentic AI deployments so it enables innovation while providing continuous, cross-application/cloud/data monitoring and real-time detection and remediation of issues (like model drift)?add
What are the key pillars of an AI operating model and how should organizations operationalize AI to get sustained value?add
>> Your employees are going to innovate. You want to give them the tools to do that. Now we know that the AI governance solutions, 80% of the governance today is periodic and manual. And that's a problem when you are in a world where you've got lots of agents acting all the time. In fact, it takes anywhere between 6 and 9 months to pick up a model drift issue or breach. How do we provide you with a hybrid solution that does governance across any applications, any cloud, any data, and also gives you the ability to tag and understand and react to alerts the minute there's a problem? So you can remediate this issue. I think culture is not something you put on walls. Culture is what you do. And so I think one of the things that attracted me to IBM is a very customer-centric culture. And so what you will see at the conference is the majority of our content is customer content, is practitioners exchanging. We built the event to be a platform for practitioners to exchange. There's a lot of what I call FOMO. I don't know if we've talked about this, but the industry talks about FOMO, the fear of missing out. Everybody wants to just experience the latest and
Dave Vellante
>> greatest.What's this new
Bruno Aziza
>> acronym?FOMU is the fear of messing up.
John Furrier
>> Oh, God. Hello, I'm John Furrier, host of theCUBE, here in New York City at IBM's One Madison. My co-host Dave Vellante. Kind of a mini CUBE Pod with CUBE alumni Bruno Aziza, Group Vice President of Enterprise Software. Bruno, you keep getting promoted and promoted. Congratulations. A lot of good stuff here at the analyst meeting laid out, got clarity around the operating model. Sovereignty was brought in, but they talked mostly international, but enterprise sovereignty comes up when you have hybrid cloud everywhere. What are you most excited about as you start to see this architecture play out and all the jewels come up from IBM's treasure chest?
Dave Vellante
>> Yes.
Bruno Aziza
>> Well, first of all, thanks for having me. It's great to spend time with both of you. We always end up with great conversations. I think the first thing I'm most excited about is the customer momentum. I think what you see in the industry is that people are moving past experimentation. They get into the real deployment issues, and sovereignty is certainly one of them. We have lots of opportunities to connect these customer stories. One such example is the upcoming conference TechXchange, which everybody should plan on attending, October 26th in Atlanta, where you'll see a lot of examples of organizations like American Express and Citi and CVS. And the idea here is that we want to bring the community together so they can orchestrate for the hybrid world. We've talked about this in the past, my belief is that the market is increasingly hybrid. Hybrid data, hybrid operations, hybrid automation. And so the complexity is not in eliminating the hybrid world. The complexity is mastering the hybrid world. And so we'll talk a lot about this at this conference. After this conference, we have Current, where we'll talk about real data, real-time data.
John Furrier
>> Well, you got a lot going on.
Bruno Aziza
>> It's another dimension of
John Furrier
>> data.as you know, we've been covering IBM for a very long time. The hybrid cloud bet was a great one that came in. The data bet, you guys have done a lot of work now with the AI-native enterprise, which is the emerging trend. Coding opened up the doors for what we're seeing as kind of an AI-native enterprise where they are having workloads move across boundaries. So security and governance start to break a little bit. Where are you seeing those two pieces? Because this becomes, at the end of the day, when people want to run agents and want to be sovereign, they want to measure that economic value, they have to nail those two things.
Dave Vellante
>> Yeah.
Bruno Aziza
>> So there's a few things going on. I think the first bit is if you're a CIO, I think 9 out of 10 CIOs say that doing AI right is going to be a career maker. At the same time, we know that the boundaryless creation is outpacing orchestration, which creates an issue here that's where governance breaks, where your employees are creating more than you can actually govern and orchestrate. I don't like the term govern. I like the idea of orchestration because it's really— you want to enable innovation and control at the same time. I think the expectation in the market is that the average Fortune 500 has 15 agents today. We expect 150,000 over the next 3 years. And so that's going to be where governance breaks, is that the amount that's created by your employees will outpace what you can manage. So you really need to build a platform that's going to allow you to catch up.
John Furrier
>> Explain what AI-native enterprise means, because I made that statement because it's not truly AI-native because they were around before AI. So they have a lot of— AI-ready, I guess, might be a better term, but they want the AI-native experience.
Bruno Aziza
>> Yeah. So there's a few things here in the AI native. It's actually interesting. I think the industry has been talking a lot about making your data AI ready. I actually think it's the other way around. It's make your AI enterprise ready that we should really be talking about, right? So where I get a little bit scared is when people start thinking about AI infused or AI as an additional capability. I think what you'll see is that most of what people are building is from AI up. With different expectations. Real-time data is one of them.
Dave Vellante
>> Let's talk about sovereignty. It's a hot topic. theCUBE Research has sort of defined 5 pillars of sovereignty. A lot of people think sovereignty is sort of narrowly defined as where you put the data, territorial. But there's so much more to it. There's certainly legal, there's operational, there's the tech stack, and the most important pillar is financial, right? We've seen people, companies token maxing. And blowing through their token budgets in a quarter that they were allocating for a year. So when you talk to leaders about sovereignty, what are they— how are they thinking about it? Do they have the full spectrum? What's your perspective?
Bruno Aziza
>> So what's great about sovereignty is it's a top initiative, right? I think 80-plus percent of organizations say sovereignty is in their top 3 priorities. The issue is that just like you noted, when people think about sovereignty, they think about data sovereignty. And data sovereignty is a good start, but often it's just location. So you have to think about data, where should it live? You have to think about the technology, which layers should be run by whom? You have to think about the operations, so who can in fact run that? And then you have to think about regulations, how are you staying in touch with that? So a little bit earlier this year, we announced a great solution called Sovereign Core that's designed to give you a complete sovereignty experience, 200+ frameworks that are allowing the organization to be in touch with regulation on a continuous basis. A great example is State Bank of India. This is a company that has close to a billion— half a billion customers. They operate in 36 countries. So they've got to manage the complexity of these regulations. So I think the answer is, of course, sovereignty. But you can't stop at data. You have to think about the multiple dimensions. And not many organizations think about that because they don't have the heritage that IBM might have on running business-critical workloads at scale across the world.
Dave Vellante
>> Would you agree with the following statement, that ultimately sovereignty is the responsibility of the end customer? A technology vendor like IBM cannot confer sovereignty to the customer. The customer has to decide what risks it takes. IBM can provide the frameworks and help the customer usher through that sovereignty journey. Do you agree with that?
Bruno Aziza
>> What I agree with is that the organization needs to have the right types of partners that are not siloing the sovereignty effort. I think what you'll see, depending on who you partner with, they might not have the breadth of solutions, so they kind of drive you down a path that is not giving you true sovereignty. True sovereignty really is across data, technology, operations, and regulation. And if you don't have a complete answer to that, you might actually go down a silo that will be detrimental to your organization.
John Furrier
>> And they want control. Control becomes the big higher-order bit.
Bruno Aziza
>> Yeah.
John Furrier
>> one thing that came up in the analyst session I want to get your thoughts on, because this is pretty obvious to us, but not, I think, to other folks outside of IBM in the industry. Identity is very strong within IBM. It's one of those things that jumps out and it's just discussed all the time. How should identity and access change when agents start talking to other agents through APIs? You start to have a lot more non-human identities. The velocity— agents work while we sleep. On our behalf in some cases, maybe not. But this is where identity— and then this brings back kind of a little bit of governance and security, but these identities are a really key piece. Explain the strategy and what you guys are doing.
Bruno Aziza
>> So there are two things that we see with customers. I think the first one, and you pointed to it, is that there's a ratio of 10 to 1 machines to humans. So by design, there is just going to be a lot more coming at your systems, more volume, more intensity. Just like you said, the agents don't sleep. The humans do, but the agents don't. The second bit is that in the way you protect your enterprise, I think 70%, close to 70% of breaches are in fact identity breaches, not password breaches. And so I think also the types of things you have to watch out for are dramatically different than when you had an identity system that was primarily designed for humans. So it really is about that, is thinking holistically about humans plus machines, knowing that the machines are going to come at you with maybe different problems than you might have anticipated in the past with humans. So we, of course, have a very broad and deep experience in those issues. I think you've seen some of the work that we've done with Vault and other solutions, but that's really the challenge. The challenge is the systems you've designed in the past just won't work for the machines.
John Furrier
>> No, I think that the operating model turns into an operating system of trust. Yeah. And you and I were talking about that yesterday when we were talking in the meeting we had in the lobby. But we hear this theme of always on.
Bruno Aziza
>> Yeah.
John Furrier
>> So continuous governance.
Bruno Aziza
>> Yeah.
John Furrier
>> The living loop, as Dave calls it.
Dave Vellante
>> Yeah, you said you don't like the term.
Bruno Aziza
>> Yeah.
Dave Vellante
>> But it's out there.
John Furrier
>> AI is a living loop.
Dave Vellante
>> And to John's point, you can't just take a snapshot. It's like security. You can't just patch once a month, right? It's got to be ongoing.
Bruno Aziza
>> What were your thoughts on that? So I didn't like the term governance in what people think about when they think about governance. They tend to think about compliance, stopping innovation, whereas I think what we want to do is We want to have a continuous system that enables people to innovate, right? like I said earlier, 150,000 agents, your employees are going to innovate. You want to give them the tools to do that. Now, we know that the average governance solutions, 80% of the governance today is periodic and manual. And that's a problem when you are in a world where you've got lots of agents acting all the time. In fact, it takes anywhere between 6 to 9 months to pick up a model drift issue or breach. And so you can tell that the old way of thinking is not going to cut it for the new world. And so that's really what we're focused on is that how do we provide you with a hybrid solution that does governance across any applications, any cloud, any data, and also gives you the ability to tag and understand and react to alerts the minute there's a problem so you can remediate this issue.
John Furrier
>> You know, one thing that came out of these big agentic experiments in production was the behavior changes of the agents themselves. They just do what they're told. If it's not scoped well, They'll say, hey, I solved your task, but I had to go through that database. I didn't have permissions for it, but I did get the answer. I got the outcome you wanted. What are you complaining about? Well, guess what? They weren't supposed to be in that database, Bruno. That's an example of the behavior that the agent makes on their own. They have their own rules if no one gives them the rules. This is the continuous change problem. It's an opportunity if you get it right. How should enterprises think about getting that right? Because that seems to be the key to the agentic system.
Bruno Aziza
>> So it goes back to what we talked about earlier, is that if you have a great container strategy and if you can write it once and run it everywhere with the same level of consistency, you win ultimately. Because you know that the world was hybrid yesterday, it's hybrid today, it's going to be more hybrid tomorrow. So I think the answer to containerizing your workloads and being able to write it once and run it in multiple places, I think it's critical. It is a little bit of an issue, I think, when we've given the choice to every developer to write wherever they want and we have not given them the tools for the enterprise to actually understand what identity, what access control is being passed across the system. So there's a lot of what I call FOMO. I don't know if we've talked about this, but industry talks about FOMO, the fear of missing out. Everybody wants to just experience the latest and greatest.
Dave Vellante
>> What's this new acronym?
Bruno Aziza
>> FOMU is the fear of messing up.
Dave Vellante
>> Oh. You got to be really careful.
John Furrier
>> I added a letter in there, FOMU. Yes.
Bruno Aziza
>> Yeah. I won't go there. But in reality, I think, the enterprise is somewhere in the middle, right? So where's the dial? Between messing up and then missing out.
Dave Vellante
>> Okay, so you got governance, identity, security, policy, agent authorization. How do these all come together in what we were talking about earlier, an operating model that a customer can deploy?
Bruno Aziza
>> So the way we have seen customers get the most value out of their initiatives is when they truly operationalize AI, right? So doing a set of experiments, getting value on the use case, and we've talked a lot about coding and customer support and marketing and sales. That's all great. Now you want to think about how you actually operationalize this continuously because it really is transformative. And so we see 4 key pillars for people to pay attention to. First one is data. And we talk a lot about real-time data. I like to talk about it as right-time data. And organizations might be OK with batch workloads and real-time, but the key is the and. You have to have batch and real-time. Then you've got to have your workloads. Every workload is agentic. Like we said, your employees are going to create agents, but the expectation of a user is now interfacing with an agent. How does that work with the nervous system of your company? How do you integrate that with the existing know-how your organization is running on? You take an example of a loan application, for instance. You already have integration for that, so how do you work with the new agentic world? Then hybrid cloud, we talked a lot about that. It's of course multiple clouds, but it's on-prem, it's at the edge, and sovereign environments. So it gets more complex to manage. And then finally, it's automation, infrastructure management. And so working across these four dimensions is what we believe is the AI operating model for IBM Software.
Dave Vellante
>> Are we at the point where that operating model, irrespective of physical location of the data and the workloads, is essentially identical?
Bruno Aziza
>> Identical, yeah.
Dave Vellante
>> Experience that a customer would see, a user would see, a developer would see, the results from the LLM across those estates? Does it look like the computer? You always say this.
Bruno Aziza
>> You're thinking about a hybrid AI OS or something like
Dave Vellante
>> that.Yes,
Bruno Aziza
>> right.You can just be an infrastructure and the application.
Dave Vellante
>> Versus, well, I kind of know I'm in the public cloud. I kind of know I'm in on-prem because it's a more secure environment. Oh, I'm at the edge.
Dave Vellante
>> Yeah.
Bruno Aziza
>> Right?So look, I'll give you my philosophy. I always think the best UI is no UI. I don't want my end user employee to think, oh, I'm running the hybrid AI OS. I just want them to do the work that they need to do. And so I think for the CIO, certainly, they're aware that they're building the building blocks for this hybrid AI OS, but I think for the end user, the UI has to be very relevant to the job to be done for that user. For you, it might be an app on the phone. For a developer, it might be a terminal somewhere.
Dave Vellante
>> But it's some kind of client interface.
John Furrier
>> Yeah.Well, you're kind of getting at the
Dave Vellante
>> headless—It acts the same irrespective of the physical location of the data and the
Bruno Aziza
>> workload.I think so. It's the headless
John Furrier
>> point That's where the headless trend gets traction because the user interface should be the best for that situation.
Bruno Aziza
>> That's right.
John Furrier
>> Not prefabricated.
Bruno Aziza
>> And that's what's amazing about AI, right? If you think about how AI is transforming how we work as human beings is that in the past, a salesperson had to work through a CRM, an application that was packaged for them where we forced a particular workflow. Now I can work with an agent that actually allows me to do the sales job. That drastically changes the productivity. But you're going to have millions of interfaces like that that are going to make humans way more effective We just need to cross that gap.
John Furrier
>> The data game is great. You got data warehouse, you got data lake house, you got data streaming. It's heterogeneous data infrastructure. It feels very infrastructure-like, Bruno. We're in a whole new era, which is why I want to ask you the question that Arvind, your CEO, just got asked by the analyst about culture. You have an event coming up, that's TechXchange, which is culturally evolving. Explain the current culture, because IBM has, decades and decades of data experience. HashiCorp brings in the Terraform, Vault, some great capabilities. You have the new products coming out that's AI native, but enterprise AI ready. So you're kind of bringing this operating model together that aligns with the computer science of the AI era. What's the new culture? What can people expect? What does TechXchange turn into as an event where people will convene and learn?
Bruno Aziza
>> So there are a few things. I think culture is not something you put on walls. Culture is what you do. And so I think one of the things that attracted me to IBM is a very customer-centric culture. And so what you will see at the conference is the majority of our content is customer content, it's practitioners exchanging. We built the event to be a platform for practitioners to exchange. Now, what are they exchanging? If you look at conferences that are available in the market, they tend to be siloed, right? So it's a cloud vendor that is going to talk about their cloud. It's an application vendor that's going to talk about their application. The reality is that you need to have capabilities in the multi-cloud, hybrid cloud, hybrid application, hybrid data, hybrid anything you want, hybrid workforce. And so the whole conference is designed around that. How do I work with IBM in this highly diverse ecosystem and what can the people that have succeeded teach each other?
John Furrier
>> It's a melting pot, but it's a tech show.
Bruno Aziza
>> Yes.
John Furrier
>> It's going to be kind of like your client zero. You bring your client zero to the table and everyone brings their—
Bruno Aziza
>> We're going to talk about what we did, but I think we're putting customers front and center. like I said, I think Citi, CVS, Yum! Brands. it's going to be an insane amount of great brands. DoorDash. And so you'll see coast to coast what these organizations have done with IBM to drive value out of it.
John Furrier
>> How much has the peer interactions changed? Because AI, it seems to us at least, it's very community oriented. People are sharing. Yeah, there's a generational set of decisions being made.
Dave Vellante
>> Yeah.
John Furrier
>> Why is it so important now for that kind of sharing?
Bruno Aziza
>> Well, it's important because the industry is moving extremely fast. We certainly see it across Confluent, HashiCorp, and you'll see this at the conference where all these communities are in fact coming together as one. The constant here is that change is permanent. And so I think people need to check with each other. It's like, I just saw this new model. I think the stat in the market is every 11 days we get a new version of a model. So it's just, it's very exhausting, I think, for a CIO or practitioner. Am I making the right choices. And they can't just turn to a vendor to say, what is the right choice? What they have to turn to is someone they know and they trust and say, how did you design
John Furrier
>> that?Yeah, we hear it all the time in theCUBE. People lock in for a week and get something built, and then they pop their head up and they go, wait, something's changed. They have to walk everything back and rebuild. Rebuilding is not good in this
Bruno Aziza
>> era.I think building for a world that changes, building for the dimensions of hybrid is the only way you can win
John Furrier
>> today.Bruno, thanks so much for coming on theCUBE in this special CUBE
Dave Vellante
>> Conversation.Great to see
John Furrier
>> you.I'm John Furrier with Dave Vellante. We are here at IBM's One Madison flagship office in New York City. Thanks for watching.
Dave Vellante
John Furrier