Insights we find valuable and hope you do too. Got something to share? Reach out.
BNI Internationals hosted an afterwork in Stockholm and I had seven minutes in front of the room. Here is what I showed them: the tools we actually run, why a new hire only ever gets crumbs of what the founder knows, and two companies who solved it from opposite starting points.
Read moreAI-onboarding means two things. Onboarding a human with AI, and onboarding an AI agent into the company. Both need exactly the same thing.
The tools split into three categories, and all three assume the same thing: that the content already exists in writing. It usually doesn't.
Who can apply, what the money actually covers, and what happens when you guess wrong. With a real worked example from Stockholm and Jönköping.
The complete method, six steps, from mapping to a source that actually stays current. Nothing held back.
You know you qualify and you know what the money can pay for. What separates a strong application from a weak one is timing, category, and paperwork, not the project itself.
Most people measure whether the source exists, not whether it works. Four real tests, and what to stop measuring.
A real template, five phases from before day one to month three, that you can copy and fill in right now. Plus what a template can never do on its own.
The binder, the PDF, the wiki, they all fail the same way. Written once, read once. Reality changes long before anyone rewrites them.
The same project can get approved or rejected depending on how it's described. A walk through the project types that actually get through, with real examples.
You've already tried ChatGPT or Claude on your business. The answers were fine, and generic. Here's why, and the four kinds of knowledge that are actually missing.
You can't always stop someone from leaving. You can stop the knowledge from leaving with them. That's an information problem, not an engagement problem.
Most people start with the most common questions. Wrong starting point. The right question is what's most expensive to lose, not what gets asked most often.
It isn't a tradeoff. The same discipline that makes a passage clear to an AI agent makes it clearer to a person too, because ambiguity is always a problem, just invisible when a colleague can answer the follow-up question.
Talks from the Future Talent Summit on AI, work, and what stays human when machines have the answers. Summarized in plain English, with the parts founders should remember pulled out.
Two days, every European SaaS founder worth knowing, and a bunch of operator talks that don't normally leave the room. Notes from the talks I caught at SaaSiest 2026 in Malmö.
The shift in sales isn't that AI made it easier. It made the system cheaper to build. Notes from the Opmore × Sakura workshop in Stockholm.
Live demos, a keynote on world models, and the first time we demoed Opmore in front of a room. What we took away from the AWS x AI Labs demo night hosted by Worker.ai.
Buyers don't pay full price for a business that runs on the founder. They cut the multiple by 30 to 50%, slow the close, and structure earn-outs that make you re-earn your own exit. The data, the diligence, and how to fix it before you go to market.
Average businesses sell for 3.5x pre-tax profit. Process-driven ones sell for 7.1x. The premium is paid for systems, not for charisma. The math, the geography, and how a €200K hire creates €2.6M in equity value.
1,300 founders. Two hours on stage. Gustaf Alströmer, Max Junestrand and Paul Graham on AI, founder dependency, Stockholm's open job, and what's actually scarce now.
A weekly summary of new insights, plus what's new at Opmore. No spam, just insights.