Tests and projects.
No client case I may show yet. Here is what I have measured and built myself.
FreeNo commitmentA yes, a no, or a plan
I test it before I sell it.
- 9 of 300
enquiries Claude Haiku 5.5 sorted wrong. All nine could be found by reading ten.
Read the test → - 215 → 92
corrections once three AI sales agents shared one memory.
Read the test → - 2.7 s
before the answer came on the call with Opus 5.5 in fast mode. Without it: 6.6 s.
Read the test →
Numbers from my own tests in the journal. Not client numbers.
Products I build and run.
Built on the same patterns I ship to clients.
PrivaThor
Removes sensitive personal data from documents before they touch an AI tool. All scanning happens offline.
Visit PrivaThor ↗AI FLUENCY GAME
The Five Moves
An AI fluency game. You learn it by doing it, not by reading about it.
See the game →KNOWLEDGE BASE
Odin
An agentic knowledge base that writes itself. Your people and your AI agents read the same, current version.
Visit Odin ↗What I build for myself.
Internal projects, where I am my own client.
Thirty minutes.
A yes, a no, or a plan.
The intake is free. I tell you whether Claude pays off in your team and what it costs, and if Brinvik is the wrong shop I say so and name a better one. You owe nothing for the half hour.
You pick a slot in Kim's calendar. No form, no callback.
Certifications from Anthropic · Fixed price on every stage, outcome agreed before the build · Invoiced in DKK