Give a workshop one clear question.

Make the question answerable in the room.

“What is the future of AI?” leaves almost every direction open. “How should we check an AI-generated explanation before using it in a lesson?” gives participants a task. The question can be debated without requiring everyone to agree on the future of an entire field.

Give everyone the same starting material.

A short text, a small demonstration or an unfinished work helps participants respond to something concrete. Send it ahead when possible. Avoid making a long presentation the price of admission to the discussion.

Ask for observations before conclusions.

What did people notice? Where did the explanation become unclear? Which part would they change? Specific observations tend to be more useful to the person doing the work than a quick verdict of “good” or “bad”. Leave room for someone to revise an initial reaction.

Design for more than the quickest speaker.

A brief period of individual writing, a smaller conversation or a visible list of questions can give different participants a way in. In a bilingual setting, a concise written summary can help more than constant word-for-word interruption.

Leave with a modest, clear record.

Keep the question, the main points of agreement, unresolved issues and any next step someone has actually accepted. The result might be one page. It should help people remember what they learned without pretending that a single conversation settled everything.