Is all the hype around AI Ops fiction or reality?
Sort by:
At this point, it's fiction. There's a lot of hype around AI. But it is not a magic bullet. It can work. But only does for the few firms that are will to put in the significant investment.
It was fiction until it becomes reality and that’s how all the hype in IT started
That’s right, for this to really work and mature, it needs time. But do we have the luxury of time in these scenarios? Because we look for instant satisfaction and instant results. We may never get it because we don’t give it enough time. We face the challenge of thinking, “We can't deliver it now so this product is junk.”
I think that's why you put an AI-Ops vendor through the pilot proof-of-concept and say, “Look, I'll give you X amount of time. Tell me how you build your model. I want to see how this proverbial sausage is made. If you can actually move the needle then we'll engage.” Time's not an issue, but it has to be mutually agreed upon. There's no magic. Too many people view AI as magic. You sprinkle AI on it and it's going to do something for you. That is not true.
Yes, we have to understand it's not magic. We have to understand that it takes time. No matter how good it is, it will take time. It's not gonna happen in 2 months.
AI-Ops has opportunity, but it’s in the extremely early innings and it's not one size fits all. You have to take a measured approach because it takes time to build proper machine learning models with data sources that are continuously changing, and you have to test this. You can't just deploy and say, “There you go, come talk to this Chatbot and let me know how it goes.” For me, one bad experience I will never talk to the Chatbot again. The culture now is, “Satisfy me now or I will not use you.”
That’s spot on. It becomes our job to understand how to bring that balance, how to get a little bit of the goodness without saying, “This is gonna solve everything under the sun.” There’s a spectrum there.
“How do I minimize my pain in IT versus how do I make my life amazing” is probably the right lens to look at it.
At the moment it is more of a hype, but It has great potential. It is a field that sees a lot of investments but still requires a lot of work to be able to automate or replace Ops functions in any significant way.
What I don’t like is seeing it as a marketing buzzword for some products, they claim to be AIOPS but in reality, they are just fancy monitoring dashboards without the ability to train the algorithm.
Some organizations are still skeptical of allowing an AI to make a decision and want AI to do the analysis and reporting/monitoring but the human to be the one that will make the final decision or action.