Transcript
The kinds of tools that you do things with, that is where you as an IT leader need an upgrade. And it's not a big upgrade. It's learning to work with GPUs in the same way that you worked with CPUs. It's learning how to mix new kinds of data and use things like RAG to be able to retrieve data for AI. It's the ability to be able to connect old systems to the new systems in a way that is seamless to the people who are using it. I believe that today's IT teams are ready for AI and they are not ready for AI because there are some things that are the same. It's compute power. Now it's new kinds of compute power, but it's compute. And it's connecting compute to data. Those things are basic. The kinds of tools that you do things with, that is where you as an IT leader need an upgrade. And it's not a big upgrade. It's learning to work with GPUs in the same way that you worked with CPUs. It's learning how to mix new kinds of data and use things like RAG to be able to retrieve data for AI. It's the ability to be able to connect old systems to the new systems in a way that is seamless to the people who are using it. But from an infrastructure standpoint, a lot of the work that you're doing is going to be the same. It's the same kind of ITIL processes. It's the same kind of building out infrastructure that is resilient, that stays up and running, that is secure, and that is enterprise ready so that when your business needs the resources, the resources are there. One of the biggest problems that I see IT teams facing today is the mix of technology and the integration that is necessary and the interoperability that is necessary to be able to run AI. Because AI requires some things that have been nice to have in the past that are absolute today. For example, today you need real-time feeds. That because AI is using real-time data to make decisions, to provide insight, to be able to act or even act autonomously as we begin. Right now we're beginning to see agents that can do things and more automation. So streaming data is more important, but it's not just streaming data by itself. It's being able to connect those data streams to the data that sits in motion and to the applications in a way that creates a closed loop. And those kinds of capabilities are not as easy as it sounds. It's one of those things we've been trying to do for a long time, and guess what? Today we have to do it. It's not easy, but it can be done. So where is AI operative today in the world of IT? Well, it's operated in the place that has been booming for the last 10 years, observability. The best thing that observability has done for the IT world is give us data across the entire set of infrastructure. And because we have that data and can make that data available to generative AI and eventually agentic AI for the sake of automating processes and procedures within the data center, we have everything we need to be able to become increasingly autonomous. Now, some areas where we're seeing an explosion of AI capabilities are some of the things that we call AI capabilities are security. Security is absolutely becoming far more autonomous. We're seeing it in terms of scalability and being able to scale up and scale down resources. We're seeing it in terms of auditability and being able to monitor in real time what is happening in broad systems. In areas where we have created this world of observability, we're pushing that observability to real time, we've been doing autonomy, and that means that generative AI and AI agents can begin to use the logic from that automation to now become even more autonomous in the way that we manage IT infrastructure. I get asked a lot about how agentic AI is useful in the IT infrastructure world and how it is different from some of the automations and scripting and kinds of things that we've used in the past. And the answer is simple. Most of what has been done to date with automations and scripting is deterministic, it's rule-based. When this happens, do this. When this happens, do that. The difference is that agentic AI operates in a probabilistic world. And instead of just taking very simplistic rules, it's able to take five or six different criteria, blend them together and determine next best action based on many different inputs. And I think that that is powerful. And the good news is there's also some fear about that agent now taking action that isn't supposed to be done. Well, the good news is that you can combine the deterministic where you have all of the rules, those rules become the guardrails for agentic AI. So you combine those two worlds and suddenly you have the safe use of AI agents in the world of IT infrastructure. So what are the barriers to adopting agentic AI? The number one barrier is fear. And that fear is actually legitimate. How do I turn over the operation of my data center to a piece of software that may be operating in a black box? And the way to overcome that fear, my recommendation is twofold. Number one, you have to jump in and begin using the agentic AI and using it with the guardrails that already exist in your more deterministic kinds of systems, your rules-based system. The second thing is your infrastructure supports a business and that infrastructure needs to operate in order for that business to run and there are different needs according to different parts of the business. So that IT infrastructure has to be well aligned with the business that it's supporting. And the best way to move forward is to have a business sponsor together with somebody in the technical side and in order for a project to actually enter the lab, it has to have business benefit attached to it and that way when it comes out of the lab, it is creating immediately business value. So how do you identify the area of your business that is most ready for agentic AI? Quite frankly, it's where you have the most data that is already creating the most value. And for me, that happens to be business intelligence and analytics and that area of your business is already being agentified and you can jump in with confidence and begin to create these conversational relationships with data, begin to create AI agent analysts that go into the data and do discovery that do the kind of work that a data analyst or a business analyst would do and send that to you as a leader. You can create a digital twin for your executives that understands the data and can jump in and begin to do the kind of work that a data analyst would do for a CEO or a chief marketing officer or a chief financial officer. That place is booming right now. I would say jump in, get going because that train has already left the station. So this whole area of AI governance is really in its infancy right now. But there's one area that we have matured and that is data governance. And the good news is that at the core of every AI implementation is data. So one place to begin with governance is make sure your data governance is in order, feed that into the AI. And then from there, the frameworks that are being developed are fairly extensive. And they include, for example, governance of the models. How do you determine which model is best for your community or for your business? And then being able to infuse that model with your code of conduct, with your code of ethics. And then being able to make sure that that model is tuned specifically to be able to do the kinds of workloads that you do within your organization. And then the security, making sure that it's secure and not only the AI, but the data is both secure. And then maybe the most important thing is auditability. AI can no longer operate in a black box. You've got to be able to know exactly what it's doing, how it's doing, and be able to reproduce how it came to the conclusion that it did and took the actions that it took. And the final one is responsibility. Because the AI has to have the same kind of responsibility that you put in place for employees in your workforce. And it may even come out of the HR department to be able to put that in place. It's a kind of a strange combination, but that kind of responsibility is incredibly important. But this is a really new area. It's just beginning to emerge. So as an IT manager, as you're preparing for what is about to hit us in the next couple of years, especially around agentic AI, my advice is upskill. Don't stay mired in the world of IT because what you'll see is you will begin to see agentic AI take over some of the manual things that have been done in IT management, IT administration, architecture, all for years. That is going to be fully agentic. So I've already said this, become the expert on how to use AI for IT and then begin to upskill. Find out right above that IT layer what are the areas where agentic AI is going to take place and become experts in those areas. And you may not be able to become an expert in agentic applications. You may not be able to become an expert in agentic analytics, but you can choose one of those and become an expert in that area so that you begin to upskill and move out of the IT management area and into that next layer that sits above it. There's also some middleware capabilities that you can learn around the way that data is moved, the way that the networks function. So don't stay in your current world, begin upskilling, and now is the time to do it. I often get asked, what is the world of ITSM going to look like five years from now? My answer is, I have no idea. We have no idea because we don't know where AI is going, but I do know this, that in that time, the world of ITSM will become increasingly invisible. It will become like an operating system where the infrastructure is there, the compute is there, the storage is there, the networks are there, but the way that they operate is completely invisible to any application, to any type of analytics, to any type of AI use case, that that infrastructure becomes more and more invisible. And so for leaders in this world, that becomes your goal. How do I use AI to make IT service management invisible so that it's like plugging in a lamp into the wall or plugging your computer into electricity? All of the things that happen on that network, you don't know about them, you don't even care about them. None of us do. We've learned that I just want to plug it in and I want it to give me whatever electricity that I need to run my application. And so ITSM will become increasingly invisible so that what you can do is figure out how do I ride that wave towards invisibility? So as a leader in the IT space, you really only have two choices. We are about to get hit by an agentic tsunami. There's been a trillion dollars invested in AI and it's moving towards agentic and it's coming and it's coming fast. Your choice is simple. Either you become the tsunami in your use of AI and agentic AI or you get tsunamied. And so you really don't have a choice. Your choice is to begin to move forward. And in that journey, my recommendation is start using AI to drive your strategy. Put together a vision for what you want to do. Use generative AI to help you fine-tune that vision. Put together a strategy for how you're going to get there. Use generative AI to be able to help you fine-tune that strategy to make sure that your strategy will achieve the vision. Use generative AI to look at your execution plans and ask it, can you remove 10 things from my execution plan that still allow me to use my strategy to achieve my vision? So become users of AI. Remember, generative AI has already read every book on ITIL, every book on ITSM, every book that will be written will be infused in it. So use it to drive your vision, your strategy, and your execution plans moving forward.
