AIOps is more important than ever—here’s why

AI integration is a priority for companies looking to improve their IT stack. Alongside that directive is a focus on AIOps and how IT operations as a whole will be positively impacted by these changes.

Bryan Krieger, CTO at RapidScale, told IT Brew that, as a rapidly evolving discipline with new challenges and responsibilities, AIOps offers leaders the chance to boost the efficiency and speed of their organizations’ AI efforts—if done right.

“It’s incumbent upon us to figure out how we best manage these new capabilities in our platforms, in our tool sets, and particularly around AI, and how we best leverage that,” Krieger said.

Must have. In an email to IT Brew, Keywords AI CEO Andy Li wrote that “AIOps is becoming a core operational capability rather than a nice-to-have.” That means more urgency around AIOps and finding the balance between operative capabilities and human oversight.

“The challenge now extends beyond monitoring infrastructure to supervising probabilistic systems that make real-time decisions, requiring centralized visibility across logs, traces, and quality, cost and latency metrics,” Li wrote. “At scale, teams need consistent ways to attribute issues across infrastructure, models, prompts, and tools, and to route the right signals to human review when automation falls short.”

Putting it down. Full organizational transparency into human/AI interaction and cooperation, as laid out by Li, is a worthy goal. But, as Runloop CEO Jonathan Wall explained, it also comes with significant challenges—primarily, deploying agentic AI solutions is making AIOps a more daunting task. It’s exciting and also offers valuable lessons when things go wrong.

“The key is, how do you deploy these things so that you can use them and that your workers are more efficient and can leverage them?” Wall said. “But also, how do you make sure that they only have access and the permissions to do the sorts of things you want them to and can’t accidentally do something catastrophic?”

A large part of the solution is digital resource management. Ensuring proper configuration is a key to a robust platform—and leads to smoother workflows. Modern devops and AIOps should work hand-in-glove, Kostas Pardalis, co-founder of data engine Typedef, told CIO Dive last year.

“DevOps is about automating and streamlining the software development life cycle,” Pardalis told the publication in November. “AIOps extends that philosophy into operations by applying machine learning and inference as first-class operations.”

Let us play. Another key to ensuring AIOps success is controlling the environment. Vanishingly few AI solutions need to be connected to the outside world; an internally restricted sandbox is often preferable for a plethora of reasons, Wall explained, not least because segmenting those operations into a container allows for more control.

Given how AIOps will frequently require access to materials that may lack the level of security preferred by most IT pros, putting agents and other AI solutions in an environment that’s relatively exfiltration-proof is a security bonus.

“This is one tool in a toolbox, and security in general is something where you build layers upon layers,” Wall said. “This is, we think, one important layer.”