Bot-Sitting - Vom KI-Werkzeug zum digitalen Team
What happens when AI stops merely answering questions and starts doing the work itself? Klaus Arent has taken that idea pretty far. He has built dozens of AI-powered projects – some useful, some profitable, some simply because he wanted to see whether they could be built. Along the way, his way of working has changed fundamentally. Today, Klaus is no longer working with just one AI assistant. Several agents can work in parallel, research, code, design, coordinate tasks and even delegate work to other agents. One agent orchestrates, others take on specialist roles. It starts to sound a lot like a real team. And that is where the **bot-sitting** begins. Because digital co-workers need management too. They ask questions. They need context. They sometimes make surprisingly good decisions – and sometimes work very efficiently in completely the wrong direction. At this cDinner, Klaus will open up his actual AI workflow and show how he works with a team of agents in practice. Not in slides, but live. We will look at how tasks are delegated, how agents coordinate with each other, where human input is still essential and how Klaus decides when to let the bots run – and when to step in. He will also build something live, giving us a hands-on look at what agentic development feels like today. The goal is not to turn everyone into an AI engineer overnight. It is to leave with a much clearer sense of **what you could already delegate to AI tomorrow**, where agent-based workflows can genuinely help, and where the human is still very much needed. Because once almost every idea becomes buildable, the real question is no longer just: **Can we build it?** It becomes: **What should we delegate – and what should we still do ourselves?**
