iConference AG
AI that lands

Nobody learns AI by watching

Your company sends its staff to an AI training session. Two hours of slides, a speaker, a handout at the end. Or the modern variant: a video series on the learning platform that everyone works through at their own pace, more or less attentively, usually while doing something else. The box gets ticked. And on Monday nobody works any differently than before.

Why slides and videos fall flat

AI is a skill, not a body of knowledge. You cannot acquire it by listening, any more than you can acquire cycling or negotiating that way. Skills grow through doing, with feedback at the moment of doing. Someone who hears a talk about good prompts knows afterwards that context matters. Someone who writes a poor prompt and gets a clarifying question back in the same breath can actually do it afterwards.

LEARNING FORMATS COMPARED Watching creates knowledge. Doing creates skill. SLIDES · TRAINING VIDEO Content Learner Passive watching One pace for everyone No link to your company No way to ask back RESULT: KNOWLEDGE DIALOGUE WITH THE AI TUTOR Tutor Learner Active doing Feedback in the moment Your own pace, your own path Work products for your company RESULT: SKILL A skill grows where the arrow points both ways.
The difference is direction: a video broadcasts. A tutor answers, asks back and adapts.

The difference between sending and answering

It does not notice when your attention drifts. It does not adjust its pace, so it bores the experienced and loses the beginners. Above all, it never asks back. Yet asking back is exactly the ability that separates an AI from a search engine, and the one your people need to master.

The tutor is the medium

In our workplace AI training, participants work through every unit as a conversation with Claude, the frontier model from Anthropic. They work with the AI on their own tasks from the very first minute. The tutor demonstrates by example instead of quizzing for definitions, and delivers something usable straight away: a draft, a structure, the one question that matters. That way participants experience during the course what good AI actually does in daily work.

Work items from your own company

From module 2 onwards, each unit produces a work product about your own organisation: a process profile, a data map, a quick-win profile. What comes out of it stays useful. At the end you hold an honest assessment of where you stand, in place of a handout.

What is verifiable at the end

Three practical questions close each unit, and all three have to be right. Complete a module and you receive a certificate. The key facts:

When you switch

Your company will build AI competence. The only open question is how. As long as the answer is slides and training videos, you are paying for attendance rather than ability. Modules 1 and 2 of our workplace AI training are free for the first 50 seats, up to two per company. Try it on yourself before you decide it for your team.

Knowledge can be presented. Ability has to be practised.

Author: Georges Leuenberger, in dialogue with Claude (Anthropic).

Read next

29 July 2026

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.

Read →
22 July 2026

The 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.

Read →
22 July 2026

Method: 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.

Read →