My AI journey started eight years ago. Not with ChatGPT, but with ML models we built on our own data — trying to improve win rates, increase deal size, or project future revenue more accurately for better decision making and operational excellence.

So I've been curious about the narrative now that “everyone is using AI.” The latest NHO report confirmed some of my thinking: 55% of Norwegian businesses say they use AI, but only 19% have moved beyond experimenting. The key skill going forward is understanding how AI connects to actual business challenges, existing processes, and what you can do with your data.

Øystein Søreide nailed it: “Det holder ikke å gi alle ansatte ChatGPT og krysse fingrene.” It is not enough to give every employee ChatGPT and cross your fingers.

Where the impact needs to happen

I don't have all the answers. But here is how I think about tying AI spend to outcomes, at three levels.

Individual — one person works faster. Measured in hours saved and output volume.

Team — workflows get smarter. Measured in forecast accuracy and cycle time.

Enterprise — AI runs on your data. Measured in win rate, churn and margin.

The tools are different at each level. So are the outcomes. Right now most conversations — and probably most spend — are concentrated at the individual level. But the KPIs that move the needle sit at team and enterprise level.

How to make it happen

Once you're clear on where the impact needs to happen, the next question is how to get there.

Optimize — turn on the AI features you're already paying for.

Buy — find specialized tools for the job.

Build — develop models on your own data.

It is tempting to build your own on your own data, but that can also take time away from what you are good at. When there are companies out there that are experts at exactly that, buying may make more sense.