Rehearsal roles
| Role | No. | What they do |
|---|---|---|
| General guide | 1 | Holds the agenda, opens the brief, delivers the perturbation (out loud and in writing, the same for everyone), calls time. Does not opine on the solutions. |
| Timekeeper | 1 | Keeps the real time log against the plan. Warns at 10 and at 5 minutes before each cutoff. Their log is an instrument of the rehearsal, not an accessory. |
| Defense panel | 3–4 | In the rehearsal: the general guide + one member of each team NOT defending that turn (rotates). Uses the judges’ sheet and the question bank. Never sums points: it records the profile. |
| Format observer | 1–2 | Does not participate: observes. Fills in the observation sheet per phase — where the format flowed, where it broke, what was missing at the table. It is the role that learns most for the real day. |
| Mentors (optional) | 0–2 | If there are extra colleagues: they guide with questions and experience, never touch the work. At the real event, mentors operate with consultation tokens equal for every team. |
With 20 guides in 4 teams, the panel rotates: when team 1 defends, one member of teams 2, 3 and 4 acts as panelist with the judges’ sheet. So everyone experiences both sides of the table — which is the point of the rehearsal.
Materials list — verify that morning
- The envelopes (or this page) with the sealed briefs — ONE is opened.
- Large paper and markers for the Human-First Window and the Divergence Gate (that phase is screen-free).
- At least one device per team (laptop or tablet; a phone works) with the connection verified BEFORE starting.
- The list of free tools verified that morning: each tool must open, without payment, without a card, from the venue’s network.
- Printed sheets: creation record (1 per team), judges’ sheet (1 per defending team × panelist), observation sheet (observers), time log (1), authoring worksheet (1 per guide, for the close or as homework).
- The chosen brief’s perturbation, printed, face down, until its moment (~1:20).
- A projector or screen for the defenses, tested beforehand.
Time log
Actual against plan, phase by phase. Deviations are data, not failures.
| Phase | Actual start | Actual end | Deviation | Note (what caused it) |
|---|---|---|---|---|
| Welcome and thesis | ||||
| Reveal | ||||
| Human-First Window | ||||
| Divergence Gate | ||||
| Build (1st half) | ||||
| Perturbation | ||||
| Build (2nd half) | ||||
| Defense prep | ||||
| Defenses | ||||
| Debrief |
Format observation sheet — per phase
For observers. One per observer; filled in the moment, not from memory.
| Phase | What flowed | What broke or confused | What was missing at the table |
|---|---|---|---|
| Welcome and thesis | |||
| Reveal | |||
| Human-First Window | |||
| Divergence Gate | |||
| Build (1st half) | |||
| Perturbation | |||
| Build (2nd half) | |||
| Defense prep | |||
| Defenses | |||
| Debrief |
Validation matrix — what the rehearsal must answer
Each question has its instrument and feeds a concrete real-day decision. When the debrief ends, this matrix is filled with what was learned and taken to the Challenge canon:
| Validation question | Instrument | Decision it feeds |
|---|---|---|
| Did the 3-hour agenda hold — or which phase asked for more time? | Time log | Real-day duration per tier (open question OQ-C19 of the canon) |
| Did the no-AI Human-First Window produce framings of the teams’ own — or were teams waiting for the screen? | Phase-3 observation + RB-01 row of the profiles | Duration and script of the Human-First Window |
| Did the Divergence Gate produce 3 real approaches — or 3 variants of one? | The divergence papers + RB-02 row | Whether the gate needs examples or more time |
| Did the perturbation stress without destroying? Was any team knocked out? | Phase-6 observation + RB-04 row | Calibration of perturbations per tier |
| Was the creation record filled in DURING — or reconstructed at the end? | The records themselves (look at ink and moment) | Record format: simpler? with triggers? |
| Did the defense distinguish the team’s own work from autopilot? Which bank question worked hardest? | Judges’ sheet + panel notes | The question bank per tier |
| Could the judges’ sheet be used live — or did it ask for training? | The filled profiles + panel complaints | Real-day judges’ briefing |
| What materials were missing at the table? | Observation sheet — missing-items section | The event’s materials list |
| Could the guides sketch briefs of their own with the authoring worksheet? | The close’s authoring worksheets | The educator preparation workshop of the real cycle |
Closing rule: what the rehearsal teaches does not stay on this page — it is recorded and proposed to the Challenge canon (its SSoT in smoother-system), which is where the format’s decisions live. This kit is an instance; the canon is the source.
Instruments prepared for the aiLearning Challenge rehearsal. Rehearsal findings are proposed to the Challenge canon — the single source of the format.