Today marks ONE YEAR since Sebastian and I started Haavn 🎉. I don't want to say we've been winging it... but in lieu of hard data around AI adoption, we've relied heavily on our instincts and experience building products. So I was struck (relieved?) to read a recent paper from INSEAD and Harvard Business School that reinforces a core insight we've used since the very beginning.
The paper studied 515 startups and found that the bottleneck in AI adoption isn't access to tools or training. It's discovery. Most companies apply AI to the obvious tasks and miss the higher-value opportunities sitting inside how they actually operate. The researchers call it "the mapping problem."
We didn't have a name for it when we started Haavn, but it's the reason we started. We both spent years in big tech building and launching products, including some of the earliest AI systems at Google. With AI in particular, people need to see what's possible first and then figure out where it fits into their work. That's a teaching and facilitation problem, not a technology problem. Surely there's a business in there??
Our very first client Renew Home started with a single workshop. The team got hands-on with AI tools fast and built their first prototypes. From there, we mapped where AI could have the biggest impact across their workflows and built six tools that automate daily operations end to end. You can't map what you don't understand, and you can't understand it from a slide deck.
Most AI adoption starts with the obvious. The real gains come from going deeper.
The funny thing about starting something new is that it becomes real the moment you say it is. How amazing! Here's to many more years of Haavn and work with Sebastian.
Paper: Mapping AI into Production: A Field Experiment on Firm Performance by Hyunjin Kim, Dahyeon Kim, and Rembrand Koning (INSEAD/HBS, March 2026)