GenAI for I&O: Changing the Perspective

Recently I blogged that while helpful, all of the various GenAI-related copilots and other add-ons being provided by I&O vendors only provide incremental benefits. I thought it only fair to expose more of my thinking and let you be the judge if I’m being just a cranky analyst that’s never seemingly satisfied or perhaps there might be some merit to my position.?

From my perspective, we’ve sliced the socio-technical system of I&O into tasks that are grouped into a functional area directed by a persona.? We may be helping that persona with these GenAI assistants, but the persona doesn’t operate in a vacuum (see DevOps as but one example where many roles may intersect).? But to current technology product and services providers, that persona is an identified buyer so they build products (or modules on top of platforms) that are specific to their needs and today, add some GenAI magic to them.? So the GenAI-oriented copilot approach fits very nicely within their go-to-market model that also leverages their installed base. Nothing wrong with this per se as it’s largely been the consumption model since the dawn of IT.

But what if some start-up, let’s call them a GenAI for I&O “native,” thinks that they can build a better mousetrap by flipping the model, i.e., starting from the data (LLM) and not the persona? ?Seems crazy wrong, maybe, but then again, it doesn’t keep reinforcing the mindset that each persona requires their own unique information or data to do their work. In fact, I think we already intrinsically know this because we’ve likely seen within our own enterprises, for example, observability and cybersecurity teams often collecting the some of the same data but through different toolsets.

So then the new paradigm that a GenAI native might be able to introduce is a breaking down of the historical I&O silos, which reduces the number of tools and which should reduce the complexity of trying to solve a problem formerly requiring the rationalization of multiple (tool) perspectives. Specifically, we may be no longer restricted to using tools designed to provide information only in the context of a specific “user case” (yes I meant to say it that way) as we already know that ChatGPT can answer a range of I&O-related questions across multiple technology domains and have it return responses in a manner appropriate to our skill level.?

Perhaps the future persona now becomes “generic” I&O or what we at Gartner have sometimes referred to as a versatilist.? The conventional wisdom used to be that an individual wearing many hats was usually a bad thing within I&O because it ultimately lead to bottlenecks.? Maybe wearing multiple hats and leveraging GenAI might be the way of the I&O future because it’s the path to reduce bottlenecks. ?Let me know what you think and here’s a collaborative research effort between industry and academia that may be showing a glimpse of this possibility: ?OWL: A Large Language Model for IT Operations.

Note: all comments or opinions expressed are mine and do not represent the views of Gartner, Inc. or its management.