Field notes on productized AI for SMBs.
What actually works in the field — diagnostic playbooks, vertical patterns, deployment war stories. Authoritative content for people building or buying AI in European SMBs.
Coming soon — first essays.
How to spot a fake AI use case in 30 minutes
A four-question test we run on every Unit candidate before scoping. Half of all stated use cases fail it.
Why monthly stocktakes are an AI-shaped hole
Reconciliation is structurally a daily problem. The monthly cadence exists because humans had no other option.
The Mac Studio that ended the BRAO argument
A field note from a Hamburg corporate boutique. Six weeks from partner meeting to live deployment.
The 4-month engagement we always refuse
Why we cap engagements at 8 weeks, and what kinds of work we send elsewhere.
Open-weight models in 2026: what's actually usable
A practitioner's view on which open-weight models we ship in production today.
Why we price like equipment, not consulting
The pricing model is the product. Here's the reasoning behind one-time, capped-scope Units.
Diagnose your top friction in 15 minutes.
One call. One operator on your side, one diagnostician on ours. We map your top three frictions to specific Units, with realistic ROI estimates — or tell you honestly that AI isn't the right answer right now.
About 1 in 4 calls ends with us recommending you don't buy anything.