H₂Tech (H₂T) sat down with Francois Laborie (FL), Executive VP at Cognite, for a Q&A about the impact of artificial intelligence (AI) and data operations on industrial operations, the importance of reliable data and the future of H₂.

H₂T: Cognite released a paper on generative AI, can you tell us a little bit about that?

FL: Generative AI is exciting because of its ability to understand large amounts of data. The question is, how does it apply to industry because we have all seen examples of discrepancies. For example, the general models or GPT model will not tell you what it does not know, it will just tell you something, and then the user must determine whether to trust it or not. So, a big part of using generative AI is understanding how to build it into a set of tools and technologies that allow you to leverage its abilities and trust the answers. We made this book available to educate our customers, and ourselves, about what this technology means for us as industrial players. What are the applications and use cases? Where do you see it being useful? Which, of course, are the ability for any operator, even if they are not a programmer to ask questions to data, potentially making code writing faster, but really making applications obsolete. Also, what are the capabilities of these technologies to look at datasets, and extract meaning out of it? All of these are what generative AI can do today, if used properly. In the book, we describe explain how to utilize these capabilities and ensure AI can be trusted to make it work for industrial players, which is much different than what we are seeing in the in the market in general.

H₂T: Does Cognite have any plans for the hydrogen (H₂) economy?

FL: We are involved in a few H₂ projects. There is a very exciting set of new technologies and potential for society in general, but there are also a lot of questions. When we speak about H₂ production, transportation and usage, ultimately, we are talking about process-based constraints, which are exactly what we are working on with the rest of our existing clients. We know that these H₂ plants must be efficient, because they from a purely economic standpoint, they must be as automated as possible. We know that transportation will be very important to design and optimize as well, and regarding consumption, we must gather a lot of knowledge and data about consumption patterns. Cognite is involved in some early phase projects and concept designs (as well as operational projects), and we are discussing how to ensure that they become more efficient and reliable.

For the full interview, visit: https://h2-tech.com/podcasts/2024/05/leveraging-generative-ai-for-industrial-advancements-with-francois-laborie/