I have been reflecting on the conversations we have been having around AI.
Just a year or two ago, many discussions seemed to be about whether or not AI could replace the service desk. It was about which roles would disappear first, and how quickly everything would change. I still hear those conversations, but definitely not as much.
Instead, we are asking different questions, good questions. Where does AI actually make sense? What should we automate, and where do we still need people? The conversations are maturing, but we still have a way to go.
When I attended SDI Spark 26, I noticed something similar. There was plenty about AI, as you would expect. But there was also leadership, culture, employee experience, and stakeholder management. It made me wonder if we're slowly moving away from talking about AI itself and spending more time talking about what changes because AI is now part of the picture.
That got me really thinking about IT Service Management (ITSM) and how we work.
We spend a surprisingly large amount of our day on work that simply has to happen. Tickets need categorising, updates need writing, and someone has to find the right knowledge article. None of that is unimportant. In fact, if it doesn't happen, service quality suffers pretty quickly.
But it's also the sort of work AI is becoming quite good at helping with. If that trend continues, I think it will be how ITSM professionals spend their day. I don't think the biggest change will be fewer tickets or faster responses.
The more I think about it, the more I come back to major incidents.
Every process describes the technical steps reasonably well, and that's the easy part to document. The harder part has always been everything around it, understanding the business impact, and knowing which stakeholders need updates first, and deciding whether restoring one service before another is worth the extra risk. Sometimes there isn't a right answer, and there are just better and worse decisions based on the situation you're in.
I don't see AI removing that. If anything, I think that kind of work becomes more visible when less of the day is spent on administration.
The same thought came back to me when I started thinking about organisations introducing AI into their ITSM processes. Getting AI into production almost feels like the obvious part. Deciding where it should be used, where people should stay involved, and when the results are actually good enough seems much harder.
It's easy to celebrate that tickets are handled faster, but much harder to notice that users are becoming frustrated, or that teams are slowly trusting recommendations they should probably question. I've already seen small examples of that.
An AI suggestion looks sensible, so it gets accepted without much thought, and often that's fine. But every now and then, the recommendation doesn't quite fit the situation, and this is where experience still matters. Knowing when to trust the recommendation is one skill, while knowing when to step back and question it is probably even more important.
Maybe that's where ITSM changes the most.
Because someone still has to keep an eye on the service as a whole, and someone still has to notice when the technology is taking the organisation in the wrong direction.
That doesn't feel like a technology problem to me, but rather like good ITSM.
I don't think frameworks suddenly become less important. Good processes still matter, and good data probably matters more than ever. AI needs something solid underneath it, or it quickly starts producing poor results.
What I do think changes is where experience becomes valuable.
If AI gives us back an hour a day, I hope we don't simply use that hour to process more tickets. I would rather see it spent understanding why those tickets exist in the first place, improving services, talking to the business, and fixing problems before they become somebody else's incident.
Maybe that's the opportunity AI gives ITSM. Not to do less thinking, but to finally have more time for it.