John Wernfeldt

About

Every CDO I talk to is pushing AI use cases, and almost none of them really trust the data underneath. I've seen a lot of AI projects fail because they never did the foundation, and the foundational work isn't really technical. It's the operating model. Who owns what, how the definitions actually get agreed, how you set up ownership so it holds. The gap I keep seeing is that nearly everything out there is tools and technique. Python, SQL, dashboards, that kind of thing. Very little on the operating model side, and that's the bit almost every organisation is missing. So people ask me which tool to buy first, or what the best business glossary is, and I say the only tool you really need is a room with the right people in it, a pen and a whiteboard. Don't start with the tool. Start with the people and the definitions. Because if you can't agree on the terms, it's super hard to take the next step. You say something in a meeting and five people think of five different things. And it's worse with AI, since AI doesn't ask which definition is the right one. It just takes one and executes. So fix one metric, one owner. I think that's the part that matters most, and the rest is very hard without it, the roles and the responsibilities and the duties and all that kind of thing. That's what I do at Northridge Analytics. Twelve years now, Gartner, Capgemini and KPMG before that, president of DAMA Sweden, writing a book on this with Wendy Batchelder for Packt. If you want an honest read on whether your foundations can carry what's being promised, I do a 30-minute call on exactly that. I'm not selling tooling. A handful each month, so DM me if you want one.