I'm not automating a response. Automating a response will still require user to read the response.
What's desired is to have save both the IT members and end users time.
Using generative AI to come up with a better process so user would not have to inquire in the first place. ServiceNow is able to collect and analyze inquiries, incidents, and requests. Feed those into generative AI to come up with an idea so there would be no inquires, incidents, and requests. It's not possible to actually eliminate them all but with each iteration of improving the procedure, it's possible to drastically reduce them.
For example, after updating chrome this morning, chrome keeps aborting on my Ubuntu 22.04. Generative AI won't come up with a reasonable solution because the update is new. It's the same with other infra questions. Over 90% of the answers are wrong because new information is not included in the model. It's actually faster to google to find somebody with similar experience and find a solution together.
What generative AI, however, can do is come up with a process so my chrome won't start aborting in the first place. I can change my automation process so abort will be less. There are ample best practices and generative AI can suggest some that would work in my particular work environment. (Most work environment has technical, human, and political constraints.)
Been doing similar with network equipments and servers as well as received inquiries, incidents, and requests.
