Generative AI hype has permeated the world of IT operations. My latest research found that 96% of IT pros who have evaluated or used generative AI for IT management use cases believe it can make people in their organizations more productive.
But attaining that productivity goal isn’t easy. First, only 41% of them believe their organizations are completely effective at evaluating the efficacy and safety or applying generative AI tools to IT operations.
We also asked research participants to identify their top barriers to getting value out of such tools. Here’s what they revealed to us:
The top response (validating quality of AI outputs) overlaps somewhat with our finding about confidence in evaluating AI tool efficacy and safety. However, validating outputs is more narrow. It’s not about figuring out what tool to buy or how to implement it safely. It’s more about knowing whether the content produced by that tool over time is good. In other words, IT pros aren’t ready to trust what generative AI tells them.
Data quality is the chief secondary issue. Here, they’re worried about the data that is used to train AI on their environment. Many organizations know that they’re IT management and monitoring tools are imperfect. Their inventory information is often out of date, and their monitoring tools often suffer errors while collecting and storing telemetry. If they forward that data to a AI vendor for training on their specific environment, the AI tool may end up having an inaccurate understanding of their infrastructure. This data quality issue may also be about creating the right prompts when interacting with the AI tool.
Third, IT pros think it’s challenging to integrate generative AI into their tools and processes. This is particularly challenging if they’re relying on general-purpose generative AI tools like ChatGPT or Google Gemini. In theory they could stream data to such tools and ask it to generate insights about their infrastructure, but that’s not what those tools are designed to do. They have general knowledge about the world. They are not capable of analyzing proprietary data from IT management systems. On the other hand, IT vendors are now offering their own AI tools to enhance their products. In these cases, the tools should be well integrated, but there is still the issue of IT processes. Any IT tool has limited value until an organization builds processes around it. This requires time to figure out how to use a tool, train people on how to integrate it into their daily operations and keep those processes up to date as time passes.
The last big roadblocks – user resistance and cost – are more mundane issues that we see whenever an organization is considering a new technology. There are always people who don’t want to change how they do things. And there are financial personnel who are resistant to adding a new line item to the budget. In both cases, IT personnel who are interesting to fostering generative AI adoption in operations need to be champions of the technology. They need to learn as much as they can about how it works and how it can deliver value. Based on the conversations I’ve had with early adopters, this isn’t hard to do with AI tools, especially if t hose tools are delivered by trusted IT vendors.
What to do next? If you’re interested in exploiting generative AI for IT operations, you should:
To learn more about how IT teams are using generative AI tools, download my market research report, or check out my free on-demand webinar.