Wednesday, July 24, 2024

How to Use AIOps in Your Business

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

AIOps is, to be sure, one of today’s leading tech buzzwords. Strictly speaking, it refers to using artificial intelligence to assist in IT operations. In a larger sense, it conjures images of leveraging AI to move your business’s technical infrastructure to an entirely new level. In this webinar, we’ll discuss:

  • What is AIOps?
  • How to get started with AIOps
  • What are the challenges and how do you work around these concerns.
  • What’s the future of AIOps? Where will we be 2-3 years from now, and how can businesses prepare now?

Please join this wide-ranging discussion with two key thought leaders in AIOps:

Ali Siddiqui, Chief Product Officer, BMC Software

Jason Bloomberg, President, Intellyx

Below are edited highlights from the discussion.

What is AIOps?


  • “So, I would say the literal meaning is ‘artificial intelligence for operations.’ And artificial intelligence is really a branch of computer science focused on machine-pattern recognition and statistical techniques to analyze their environment, so in the AIOps case it would be analyzing the IT operations environment.”
  • “And just to give context, I would say, going forward more and more, AIOps will be about being able to see things before they happen out in the horizon. Now I know that’s Nirvana, it sounds a little bit too much, but really this is a key technology which really moves forward the needle towards digital transformation for ITOps.”


  • “Well, it’s actually, for better or worse, a Gartner term, and Gartner loves inventing terms, and sometimes they stick, a lot of times they don’t, they just make them up all the time. And it originally meant algorithmic infrastructure op for operations, and nobody had a clue what they meant by that, and it just didn’t make any sense at all.”
  • “But it didn’t catch on, so they quickly reversed themselves and said, ‘Well, AI, everybody thinks it means artificial intelligence, so that’s really what we meant all along, and it’s artificial intelligence for operations.’ But as such, it is really Gartner’s effort to take these two things, we have the IT operations management or ITOM, and we have artificial intelligence, AI, and it’s like, ‘Well, these fit together somehow, so let’s stick them together and see what the vendors do about it.'”

JM: So is AIOps primarily for IT? Or is AIOps this larger thing that helps us run the entire enterprise?

Bloomberg: Well, I would say that that may be the case indirectly, given that today’s modern digital enterprise is software-empowered. So if you look at a digitally transformed organization, software helps them connect to customers across the organization. And as such, making sure that software is working properly and is robust and meets the business need is a core part of now the business priority, the business now cares as much as IT cares. So in that sense, yeah, AIOps is going to be supporting this vision of the digital enterprise, but it’s still really about IT operations.”

If a business wants to get on board with AIOps, what are some first steps?

Siddiqui: “So at the IT organization, you need to pick a few of these use cases, one or two ideally, and get your team to  the right vendor who addresses those specific use cases and really have some major goals that tell you whether you’re successful or not. So that’s one of the key things I would say you need to do to start with, not just assessment, but also know: where do you want to end up? What is the success criteria?”

Maguire: What might be a tangible, real-world example, Ali, of how this would work?

Siddiqui: “So if you look at, from an ITOps perspective, anomaly detection would be one of the use cases. What that really means from a use case perspective is you don’t want to go manually set thresholds or do baselining, which is not that accurate. You want to reduce noise by system learning, you had said ‘something that self-learns and self-corrects and then fixes issue.’ System learning and doing anomaly detection of issues, and then telling you, not just anomaly detection.”


  • Well, it’s important to understand that in the enterprise ops context, there are already many different tools, some of them are older, some of them might be more recent, if anything, there are too many tools. And this is part of the challenge is that we have all these different tools.”
  • Remember, from the CIO’s perspective, not only do they have a lot of tools, but they’ve already spent a lot of money on a lot of tools, so there has to be a real cost justification for bringing in a new tool. So there’s really two approaches, one is that we need to retire one or more older tools because they no longer meet the business need, or they no longer support it.”
  • “Now, the second part, the filling the gap is perhaps the most common approach to bringing AIOps into the organization, and the gap that is often the problem is the problem of too many events. Older ops technology will generate events all the time, those events could represent real problems or they could just represent no problem at all, or some minor problem, or it could be multiple notices about the same problem, and then you end up with this cry wolf situation where the ops person is seeing dozens of events every hour, every single hour, every single day, 24/7.”
  • “AIOps is especially good at solving that problem, it’s especially good at identifying those anomalies, so those singular events that you need to pay attention to. It’s good at reducing the effect of the less important events. And AIOps is also good at combining related events…those are now extra benefits once you have the AIOps in there to help with your event storms.”

What are some challenges with deploying AIOps, and how do companies work with these challenges?

Siddiqui: “I would say that the challenge really is to not forget that it’s a three-legged stool. Tools is just one part vendors like us provide, and there are some really good tools from across the vendor base right now. But people and process, it’s a three-legged stool. So as enterprises go on this journey of AIOps, they have to think of all three: People, process, and tools.”

  • “In the end, it’s really about reinventing your business as a technology-powered business. So the focus really is agility, customer-centricity.”
  • “My point was that in the end the business outcome has to be paramount, whether it’s agility, whether it’s customer experience, improving your customer experience, so that has to be paramount.”


  • “It’s important to understand that AIOps and the use of AI in the context of AIOps is related to but different from the use of AI in the enterprise today, generally speaking. AI is particularly useful in the enterprise for extracting insights in large data sets. And that could be for any enterprise that deals with large data sets as part of their day-to-day business.”
  • “So this is where a lot of the AIOps vendors are stepping in. They’re saying, ‘Well, we’ll give you the AIOps technology where you don’t need to assign a data scientist to deal with the AIOps technology.’ You assign the operators, the SRE’s, and other people who are dealing with the operational environment. And in many cases, the machine learning or the AI part of AIOps is essentially it takes care of itself.”
  • “The other part of the story is the security part of the AIOps story. The anomalies that AIOps can detect, they could just be that some system’s running out of memory or something, some sort of operational anomaly. But more often than not, it’s some sort of cybersecurity anomaly. It’s a sign there’s some attack going on.”

What is the future of AIOps and how can companies prepare for this future?

Siddiqui: “So if you look at the future of AIOps, I think in the next few years, there will be some pull coming. There is pull coming from too much data. It’s not humanly possible to look at this deluge of data and make analysis. So you need out-of-the-box machine learning that can give you insight from this deluge of data, whether it’s logs, events, metrics, it doesn’t matter. Then the other thing is the strengthening of IT Ops and dev ops, that becoming even more closer together, the close tie-in will also drive some the AIOps needs.”

  • “So there has to be clear ROI, number one, for the organization, and tool consolidation could be improving customer experience. So AIOps not just helps in operational efficiencies, but helps in your digital transformation. So IT Ops needs to work with line of business if AIOps is to fulfill the full height.”
  • “So if I had to look at the next few years, I’m optimistic because of these reasons, that the autonomous digital enterprise journey, the line of business, and IT Ops working closely together to figure out the business outcome. Focusing on that, the vendors like ourselves [are] making tools that are easy to use. The machine learning does not require a data scientist out-of-box.”
  • “We have many customers, and many of them are still hybrid, which means they’re moving to the cloud, the journey is not over. And that means there is on-prem requirements. And even if I look at our BMC AMI AIOps suite, specifically the AMI Operational Insight product that we just launched, well, it does work on on-prem.”

Bloomberg: “Well, the only reason we’re talking about AIOps now is because AI is the new kid on the block. Give it a few years, and all ops management will be AI-empowered. And if everything is AI-empowered, then AIOps will just be ops management, you stop thinking it would be a distinct category.”

  • “But there’s a couple of primary drivers for why this is going to happen in the enterprise infrastructure environment. One is that when we look at the broad scope of what’s going on, we’re moving to a model I like to call cloud native computing, where we’re leveraging the best practices of the cloud for all of enterprise IT, including hybrid IT.”
  • “But secondly, this environment is also extraordinarily dynamic. This is one of the capabilities that the world of containers, the Kubernetes world, brings to traditional virtual machines is an extraordinarily dynamic enterprise infrastructure environment. Well, when you have a very dynamic environment that is generating vast quantities of data, there’s simply no way to manage that kind of environment manually, you have to automate it. All of those trends are not going to stop, and AIOps is gonna be increasingly essential, and before long, it’ll be ubiquitous, we’re not gonna have anything that isn’t AIOps anymore.”

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles