In the Workshop

A workshop with MEADOWS is not a lecture about AI tools. It's a shared space where teachers use AI tools together, in real time, and learn from watching each other work.

Peer learning in a shared space

Teachers collaborate around the same bots, in the same conversation. When one does something clever with an AI — a prompt, a workflow, a summarisation move — the others see it happen and adopt it. AI facilitation skill spreads horizontally, by demonstration, not by course.

This is the interactional labeling principle applied to professional development: the "nice move" moments are visible and shareable, not buried in individual practice.

Teachers as authors

Educators build their own small bots tailored to their subject — a Socratic questioner, a rubric-checker, a debate partner — without needing to be engineers. Ownership of tooling stays with the teacher. The docent test enforces this: every SDK surface must be usable by a teacher working with an AI during a hackathon.

# A Socratic questioner — teacher-authored, no engineering degree required
from meadows.bot import BaseBot

class SocratesBot(BaseBot):
    BOT_NAME = "socrates"
    BOT_DESCRIPTION = "Asks probing questions about whatever you discuss"

    def should_handle(self, command, args):
        return command == "ask"

    def handle(self, command, args, raw_args, message, thread_context):
        topic = " ".join(args) if args else "that"
        return f"What do you mean by {topic}? Why do you think that?"

if __name__ == "__main__":
    SocratesBot().connect()

Mining the archive

Because sessions persist as append-only JSONL, a workshop can look back at what previous groups discovered, reuse patterns, and stand on prior work rather than restarting from zero.

What teachers actually learn

Concept	Where it shows up
Bot authoring	BaseBot contract: BOT_NAME + should_handle + handle + connect
AI facilitation	Watching peers use LLM bots, prompts, and forms in the shared conversation
Label-based routing	Subscribing to labels to filter what a bot sees
Interactive forms	send_form() for structured input
Systems thinking	Observing how bots interact, cascade, and produce emergent behavior

Check your understanding

• How does a teacher author a bot without being an engineer?
• What does "append-only persistence" mean for workshop archives?
• Why is the shared conversation more effective than individual practice for learning AI facilitation?

See the Concepts page or jump to the first bot tutorial.