Session details

Your cohort’s join link, date, and time appear here once the schedule is confirmed. After the session, the recording will be posted here too — along with office hours, if your cohort has them.

The session runs up to three hours, including a break.

Before you join

Everything you need is listed in Studio 4 Overview. One addition: be ready to share your screen.

How the session runs

1. Orientation and team setup.
2. A first build round, with light structure.
3. Breakout discussion, then whole-group debrief — what happened?
4. Roles and rotation introduced.
5. A second build round, worked differently.
6. A break.
7. A third round: something changes, and your team has to recover.
8. Your team writes its Collaboration Protocol.
9. Debrief, reflection, and your field experiment.

Roles rotate on a five-minute cadence. The facilitator runs the clock — you do not need to watch it.

Your team setup

• Four people per team.
• One shared AI conversation for the whole team — not four private ones.
• One shared notes document.
• One person sharing screen at a time.
• A way to run or inspect tests.

The opening exercise

A team task board marks work items as stale, blocked, or ready for review, based on status, age, assignment, and verification state.

Use AI to add stale-item behavior to the task board. A work item is stale after five working days without activity. Make the tests pass. You have fifteen minutes.

Working with AI during the session

A few things worth knowing, because they will come up:

• AI can be confidently wrong. It has no reliable sense of its own uncertainty, and it sounds the same whether it is right or guessing. Fluency is not evidence. This is the most important thing to hold onto today — verification is not bureaucracy, it is the mechanism that catches what confidence hides.
• Give it the why, not just the what. AI working from intent behaves noticeably differently from AI working from a task. It cannot infer what matters to your team.
• Ask for the smallest next step rather than the whole solution. Large answers are harder to check, and unchecked answers are where drift begins.
• It will not tell you to stop. AI has momentum and no instinct to back up. Deciding an approach is wrong remains a human move.

During the session

The feature is not the point. It is there to create pressure. What we are studying is whether a team keeps shared understanding alive while AI participates in the work.

If you are mid-thought when the rotation timer goes off, capture it in the shared notes and rotate anyway.