Our position on AI for mid-sized companies
Plain talk on AI in the Swiss SME world: what actually makes a difference, without the noise.
Private AI accounts at work: ban first, then introduce properly
Customer data in a private AI account: who is liable under the revised Swiss data protection act, how thin the case law still is, and the two steps an SME takes now.
ReadThe most insidious trap in AI answers: they sound utterly convincing but are factually wrong
We tested four AI models on a fictitious contract file. The most dangerous result was not the error, but how convincingly the weakest model went wrong.
ReadMethod: four AI models on an M&A contract file
Method and findings for the model test: four AI models against a fictitious, anonymised contract file, measured deterministically against a solution key.
ReadIn the beginning there was fire
Jürgen Schmidhuber compares AI to fire: useful and dangerous at once. Why companies should bring the flame indoors under control instead of waiting.
ReadNobody learns AI by watching
AI is a skill, not a body of knowledge. Slides and training videos convey knowledge, but not skill. Why AI competence in a company grows through dialogue.
ReadAI rarely fails on the technology. It usually fails on the foundation.
Gartner estimates over 40 percent of agentic AI projects will be scrapped by end of 2027 — the cause is rarely the tool. Four pillars decide whether AI carries in your business: processes, data, systems, ground rules.
ReadLocal LLMs are no substitute for frontier. They are the other half.
Open-source models lag frontier by just «four months»? For an SME in a competitive market, that's the wrong conclusion. Why sovereignty isn't capability — and what the right mix of frontier and local models looks like.
ReadAI on the phone — and it understands Bernese German.
We tested an AI phone assistant: it understands genuine Swiss German, runs the conversation cleanly and stays composed on follow-up questions. An experience report — and the start of a pilot.
Read«What» gets you what you asked for. «Why» gets you what you need.
On big AI projects, the first prompt sets the course. Why «How» belongs in the guardrails, «What» just gets executed — and only «Why» makes the AI think along.
ReadChatGPT, Claude and Gemini at work: allowed — but not in every version.
The problem is rarely the tool — it's the version. What a business licence and a contract really change, why the DPF list isn't the exclusion criterion, and what an SME should do in practice. Plain talk, not legal advice.
ReadAI meets SharePoint: a pain to set up, a win to run.
A company-wide AI layer on Microsoft 365 — Claude via MCP to SharePoint, email and Teams. An experience report: setup is tedious, running it convinces, and on balance it pays off.
ReadSaying «Claude, Codex or Gemini Pro» today gets you the tool. Not yet the solution.
Your best caseworker knows every file. Imagine everyone in your company could do the same — grounded in your curated company knowledge.
Read