Agents do the work. It still doesn’t look AI-made.

Everyone can get agents to produce something now. The problem is what comes out. The text reads like a machine wrote it, the page looks like ten thousand others, and the reader has decided you did not care before they judged a single argument. The usual answer is to write more prompts. That is not where the fix is.

I build websites and applications for Swiss clients, AI agents do most of the work, and you cannot tell from the results. In these 90 minutes I show that team: which instance decides, which ones only execute, where a different model gets pulled in to argue, and which decisions I never hand over at all. Then the part nobody talks about, the gates that catch the machine default before a client ever sees it, and what each one cost me to learn.

You will not work on your own material. This is a demonstration with honest advice, not a workshop. No camera needed, listening and asking in the chat is enough. I will also tell you which parts of this are worth building for your own team and which are not worth the maintenance.

Who it is for
Anyone shipping work with AI where the result gets judged: independents, marketing and communications, design, development, agency people. No coding needed.

Format
Online, a talk with a live demonstration, 90 minutes, questions in the chat throughout. No recording.