What is AI – REALLY? Fun? Useful? Revolutionary? Statistics? Scary?

We are surrounded by AI, it is in our phones, computers, vacuums and even fridges. At this point we may even expect AI to be in “everything”. But, have you ever taken a step back and thought about what AI really is and what value it brings you? In essence AI is merely a combination of functions and weights, and its output, although often impressive, nothing more than the answer resulting from a large number of calculations based on the input provided. One may even argue that this is similar to how we as humans arrive to conclusions.

While generative AI (GenAI) has been in the “spotlight” for less than two years, AI as a whole has been around for quite a while (actually since the 1950s!). Experience has shown that AI time and time again falls short of enabling change that is more than “skin-deep”. Applying AI to the right problems and in the right way is integral to sustainable value creation.

This week Microsoft and Startuplab put on a seminar with introduction to AI and “copilots” at Forskningsparken (Oslo Science Park). The day started with high-level general introductions before diving deeper down into the building blocks of Microsoft currently AI offering. While in many ways very impressive, the potential impact still feels “skin-deep”, as if we now have access to fire but are still trying to figure out how to use it constructively.

We believe that domain expertise is the key to going deeper, and we were lucky enough to share our thoughts at the end of the seminar, alongside excellent presentation from novem and Masterchannel. By combining AI with a domain-informed data model and experience-based best practices we can remove the “co” from “copilot”, providing (auto)pilots for the commercial management of renewable power generation plants.

With the flexibility and creativity of AI, and the rigidity of formalised domain expertise we can create the reliability and predictability required to go beyond “skin-deep” applications, enabling realisation of sustainable value. In many way you could argue that combining domain knowledge with AI is like bowling with bumpers – there is still room to move, but you’re protected from the “gutter”.

We’re curious, what is your experience with using AI?

Check out the LinkedIn post here.