RESOLVE ’22: Expert predictions for AIOps 2022-2025
BigPanda’s RESOLVE ‘22 conference hosted a number of luminaries in the AIOps and IT Ops world, so naturally we needed to get their thoughts on the future of the market and where they see AIOps going in the next few years.
Our guests for the session titled Expert predictions for AIOps 2022-2025 were from the press, investor community, analyst community and vendor world. Chief Product Officer at BigPanda Fred Koopmans kicked things off with unique insights from James Maguire, editor in chief at eWeek; Carlos Casanova, principal analyst at Forrester; Eric Noeth, partner at Advent; and Michael Yamnitsky, managing director at Insight Partners.
To start, Fred asked our guests to explain why they think AIOps is where it is today.
How did AIOps get here?
Today’s enterprise IT systems are just too complex for humans to run,” James said. “In the old days, there was a thing called a data center, and IT pros ran it. If something went down, they could put a piece of tape on the server—maybe talk nicely to it. That world is gone and is never coming back,” he added.
Part of the evolution of AIOps is simply how fast enterprise IT systems move. With cloud, edge, microservices, Kubernetes, massive levels of data and more, James noted that the level of data is just too much for humans. Plain and simple, they need AIOps to help them handle it. For James, there is irony in AIOps’ evolution: “We created these systems that are so complex that we ourselves can no longer manage them. So we created a second system called AIOps to manage the first, complex system.”
Carlos added that CI/CD pipelines and “all sorts of other methodologies and philosophies are forcing operations” to transform, “start stepping up their game” and maintain new ways of operating.
Eric introduced another concept to the stack behind why AIOps is having its moment: the scarcity of IT talent today. Not only has there been “an explosion of innovation up and down the IT stack,” as he said, but there is a “market backdrop” of a lack of skilled IT pros making matters worse. The demand is high for these IT people—especially given the amount of innovation happening—as problems get more complicated and the stakes rise for businesses. AIOps is proving itself the answer to this skill shortage problem.
The last trend offered was from Michael who noted that it’s getting increasingly more important for enterprises to be able to see everything going on in their IT systems. AIOps offers the ability to view in real time what’s going on in your system through one single pane of glass.
What do the next couple of years hold for AIOps?
In terms of what’s in the short-term future for AIOps—but hasn’t fully arrived—Michael said “we are going to see a real improvement in terms of building the ultimate service map of all the different relationships across infrastructure and applications.” We will begin to be able to see how pieces of the IT environment relate to one another as they degrade so we can have a bigger picture of both application performance and business impact.
He also mentioned that a trend he sees is “way more insights and capabilities through the DevOps teams and engineering teams that are actually responsible for fixing problems.” Not just AIOps best serving the L1 folks.
And businesses are tracking the financial or brand impact of downtime more closely today as well. Carlos noted this is going to improve in accuracy through AIOps. Organizations “can’t put a true dollar number on it if they don’t know what those services fully look like,” he said—a problem that can be tackled if they can have a full-service map with “costing at each tier of microservice.”
For Eric, the opportunities that will emerge soon are simply around the continued possibilities for automation—especially on the IT side of the house. And for James, the ability to predict will also grow AIOps. “We need predictive analytics using maps and data sets to forecast the future, which AIOps can do,” he said. “In fact, we have a desperate need for that.”
“Your automation strategy is only as good as your data strategy.”
That issue—and others—are holding enterprises back from fully embracing automation and AIOps. As Carlos put it, “organizations are just not ready to turn over the keys.” Without the vital factors of “trust and transparency,” organizations will continue to lag in automation adoption. Why? Carlos explains that even though they compare what a machine does to what a human does over and over again—clearly seeing the results in favor of the machine—there is mental resistance against turning over those keys to AI. Carlos suspects that part of the reason is just human wariness of artificial intelligence.
But, as Michael pointed out, “we’ve reached a point where it’s not about the robots taking [our] job[s]. It’s about the robots being there for our survival,” so IT Ops folks can finish their days and go be with their families. There is no way to move forward without AI as a part of the IT scheme from a sustainable viewpoint, and Michael would “love to see things improve where we as a society just accept that AI is embedded in everything.”
Eric offered that people are starting to “get it.” Despite persistent fears, organizations are realizing that AIOps can solve a real problem—especially as “every vendor that touches AIOps in some way is one talking about AIOps.”
The biggest accelerator of AIOps? The pandemic.
Carlos said that the acceleration of AIOps over the last five years is really because of the events of 2020. Without the pandemic, it likely would have taken a while longer for AIOps to see the growth it’s seen. James agreed, but noted, “[The pandemic] didn’t create the change, but without it, I think we would have gotten to [today’s] point by 2024.”
The panelists offered several AIOps trends coming down the pipe:
- Workloads will move to lower-cost platforms to act preventatively or cost-efficiently. As Fred put it, this will help not to prevent the issue from happening, but to prevent “the manual intervention required to fix it.”
- Organizations will use AIOps to evaluate their costs as it enables them to build a map of their entire infrastructure. Michael emphasized that cross-domain insights will “be coming out of the array of IT Ops platforms.”
- Security operations center (SOC) and network operations center (NOC) will converge, and AIOps tools in the market will get simpler—according to Eric.
James challenged Eric on that last point, noting that AIOps is actually going to get more complex. But Eric explained that while “everything behind the scenes is going to get immensely more complex,” the “actual user interface and how it’s implemented out of the gate” will be barriers to adoption that need to come down over time. That’s where James pointed out that the competitive advantage in AIOps will go to the provider that can offer a “low-code/no-code approach to AIOps.”
AIOps for all
James and Carlos discussed the need to abstract the AIOps concept—to speak less about the technology and “more about the outcomes”—so that AIOps is more accessible to less techy people. More people will be touching AIOps in the future, and they won’t all be AI wizards. James referred to this concept as the democratization of technology. Eventually, AIOps will serve more organizations and folks in different roles at those organizations. This underscores Eric’s previous point about the simplification of the user interface. As more people are able to get involved and have a hand in AIOps, it will need to be easier to use and its outcomes more clear.
“AIOps offers a multi-layered tool that allows a small, nimble team to compete with much larger, well-funded companies,” James said. “In many ways, I think AIOps will be a major step forward to the democratization of technology.”
Watch the full session on-demand here.