IT operations teams already automate many tasks, such as performing software updates and generating alerts when anomalies appear within monitoring data. But there are still some workflows within ITOps, like incident response and reading incoming tickets, that traditionally require a lot of manual effort.
Could generative AI be the key to streamlining those tasks and bringing ITOps automation to the next level? You wouldn't be crazy for thinking so. In some respects, generative AI for ITOps is poised to be the next big efficiency gain for IT organizations.
On the other hand, it's important not to overstate the potential of generative AI in ITOps. The technology offers promise, but it probably won't change every aspect of IT operations.
Keep reading for a balanced take on what generative AI for ITOps can and can't do.
- IT Operations Challenges That Generative AI Can Address
- The Limitations of Generative AI for ITOps
- Where Are the GenAI Tools for ITOps?
IT Operations Challenges That Generative AI Can Address
Again, the key opportunities for integrating generative AI into ITOps surround IT operations processes that engineers have traditionally struggled to automate. Those processes tend to be ones where conditions are highly inconsistent or unpredictable, making it necessary to have humans perform nuanced evaluations to make decisions.
For example, deciding how to respond to an incident like a server failure is not a task that ITOps teams can typically outsource to an algorithm. They might use automated tools to help understand why the server failed and determine whether the outage has impacted other resources, but planning and implementing the process necessary to resolve the outage is a task that humans have to handle manually.
Similarly, interpreting incoming tickets in your help desk can be tough to automate fully. Methods such as keyword and sentiment analysis — which are examples of analytics techniques that don't require generative AI — can help help desk teams sort tickets automatically based on type of request or urgency. But they do little to streamline the task of actually reading a ticket and figuring out what a user really wants. You need a human to do that — or you did, at least, until generative AI came along.
The ability of generative AI technology to interpret complex situations on a nuanced, case-by-case basis means that generative AI may be capable of solving challenges that other approaches — including traditional AI/ML-based pattern matching — couldn't handle. It's not hard to imagine generative AI performing tasks like the following:
- Summarizing tickets, or condensing multiple requests into a single ticket, so that IT engineers can read them faster.
- Automatically generating playbooks tailored to individual instances to guide response efforts.
- Helping IT engineers locate information more quickly by summarizing documentation or making it possible to pose technical questions in natural language and receive answers based on an organization's knowledge base.
- Generating code for maintenance scripts, infrastructure-as-code templates, and other code-based resources that IT engineers have to produce when setting up automations.
This certainly isn't an exhaustive list, but it provides a sense of how generative AI can potentially speed up and simplify tasks that would take IT engineers a long time to perform manually.
The Limitations of Generative AI for ITOps
That said, don't expect generative AI to revolutionize ITOps or usher in the age of NoOps (meaning a world where ITOps processes are so completely automated that there's no need for humans to remain in the loop at all).
Applying generative AI to tasks like those described above will save some time and effort, but it won't fully automate complex workflows. For example, the ability to summarize and condense help desk tickets might make the task of reading tickets perhaps 20% or 30% faster, but it won't automate the work of figuring out which actions to take to resolve a ticket — work that accounts for the majority of the time that help desk staff spend handling a request.
Likewise, auto-generated incident response playbooks may help response teams operate more efficiently. But because the playbooks might contain errors or oversights, there will still be a significant level of human involvement necessary to interpret the playbooks, fill in the blanks, and oversee response efforts.
The point here is that, while generative AI may certainly add some efficiency to the way ITOps engineers work, it won't double or triple their overall efficiency. The gains will be incremental.
Where Are the GenAI Tools for ITOps?
It's worth noting, too, that actually leveraging generative AI in ITOps will require teams to have tools at their disposal that offer generative AI features — and so far, those tools are in short supply.
Some vendors — like Rezolve.ai and Forethought — offer solutions that leverage generative AI to streamline some ITOps tasks. But they focus mostly on providing chat interfaces linked to automations that can solve basic user requests. Vendors have yet to build out a broader set of generative AI-powered features targeted at diverse ITOps needs.
Until (and unless) the tooling becomes more mature, the ability of ITOps teams to take advantage of generative AI at present will be limited.
Conclusion: GenAI Won't Transform ITOps Work
ITOps teams looking for ways to add efficiency to basic processes, like ticket management, can benefit from solutions that incorporate generative AI today. There is also potential for generative AI to streamline a variety of other IT operations tasks, but currently few tools cater to these use cases. And even if they did, most GenAI-based features for ITOps will only move the needle incrementally, rather than revolutionizing the way ITOps teams work.
About the author
Christopher Tozzi is a technology analyst with subject matter expertise in cloud computing, application development, open source software, virtualization, containers and more. He also lectures at a major university in the Albany, New York, area. His book, “For Fun and Profit: A History of the Free and Open Source Software Revolution,” was published by MIT Press.
