ISTE AI Pavilion — Flash Lab Table-Card Scaffold
For: Wed Jul 1 Flash Lab session, ISTE AI Pavilion (noisy, station-style room) · Reuben + Josh Design thesis: In a loud pavilion your voice won’t carry, so the card is the facilitator. Pre-framed topic + printed boosts + 3-person groups = the structure runs the table; you and Josh rove and nudge.
The Flash Lab mechanic (the thing that makes it Flash Lab): Three areas — Design, Build, Advocate. In each area the group picks one boost to get the work done: a 🧑 human boost (an activity you do with other people) or a 🤖 AI boost (an activity you do with AI). Choosing — and noticing the pattern of your choices across the three areas — is the experience. That’s how a participant makes an informed decision about when they want AI and when they want a human.
Directly answers Isabelle’s DCI feedback:
- #1 Topic framing was murky (NY handshake) → topic is pre-set (AI-Ready Graduate) and the problem statement is pre-printed. No open brainstorm to get lost in.
- #2 Groups too big / one voice dominates → 3 per table, hard cap. Every person makes their own AI-or-human call, and the human boosts need more than one set of hands — no room for one person to run the table.
- Josh’s “do we hire a facilitator per table?” → No. The card carries the structure; you two rove.
THE TABLE CARD (master template — one per table, role slotted in)
Print double-sided, large type. Each table gets this card with one role’s problem statement (below) dropped into the top box.
PORTRAIT OF AN AI-READY GRADUATE — FLASH LAB Table focus ▶ [ROLE]
Your group of 3 moves through 3 areas: DESIGN → BUILD → ADVOCATE.
In each area, pick ONE boost to get the work done:
🧑 a HUMAN boost (do it with other people) or
🤖 an AI boost (do it with AI)
Picking is the point. Notice when you reach for AI and when you reach
for a human — that's your call to make.
──────────────────────────────────────────────────────────────────────
YOUR PROBLEM (read together · 3 min)
[PRE-PRINTED ROLE PROBLEM STATEMENT — see set below]
① DESIGN — sharpen the problem & who it's for ~14 min
Pick ONE boost, then do it:
🧑 Borrow an outsider — pull someone from another table. They
grill your problem with questions until it sharpens. You only
answer; no pitching solutions.
🤖 Question machine — paste: "Ask me 5 questions to find what I'm
missing about this problem. Don't suggest solutions yet."
▸ Why did you pick AI or a human for this one?
② BUILD — make a rough version of your idea ~22 min ← meat
Pick ONE boost, then do it:
🧑 Co-sketch — build it on paper WITH a partner, yes-and-ing each
other. Two hands, one prototype.
🤖 Draft three — have AI generate 3 rough versions; you cut,
keep, and remix the parts that work.
▸ Why did you pick AI or a human for this one?
③ ADVOCATE — make the case for it ~14 min
Pick ONE boost, then do it:
🧑 Pitch to a face — give your 30-sec pitch to a real person.
Ask for their gut reaction + one hard question.
🤖 Hostile reviewer — paste: "Be a skeptical school-board member.
Poke holes in my pitch and ask me the hard questions."
▸ Why did you pick AI or a human for this one?
LOOK AT YOUR THREE CHOICES (~6 min)
Where did you reach for AI? Where for a human? What does your
pattern tell you about when AI helps you — and when it doesn't?
Sticky on the board: "I chose AI for ___ , a human for ___ ,
because ___ ."
──────────────────────────────────────────────────────────────────────
That's the 60-minute core. Still buzzing? Flip ▶ for how to run the
whole Flash Lab — and bring it to your school.
Run of Show (90-min slot · meat in the first 60)
Front-load the value so anyone who drifts at the hour still made all three choices and saw their pattern.
| Time | Block | What |
|---|---|---|
| 0:00–0:05 | Welcome + frame | “You’ll make three real choices today: AI or a human, for each part of the work.” Point to the Human-or-AI board. |
| 0:05–1:00 | The table card (60-min core) | Design → Build → Advocate, each a boost choice, then the reflection. The full experience lands inside the hour. |
| 1:00–1:20 | Cross-table share + the pitch | Gallery-walk the “I chose AI for… / a human for…” stickies; you + Josh make the two CTAs (below). |
| 1:20–1:30 | Capture + linger | QR sign-up; talk to whoever stayed — those are your hottest leads. |
STEP-0 PROBLEM STATEMENTS (one per table, by Profile role)
Each frames a real teaching problem the group will Design / Build / Advocate around. The AI-or-human choice plays out inside it:
- Learner — “How might we help students learn how to learn in an AI world — without short-circuiting the productive struggle that builds understanding?”
- Problem-Solver — “How might we help students decide where AI carries a problem and where they need to own the reasoning themselves?”
- Synthesizer — “How might we help students pull scattered sources into one coherent idea — keeping the meaning-making theirs?”
- Connector — “How might we use AI to strengthen how students connect ideas and people — without replacing the connections that matter?”
- Storyteller — “How might we help students tell a truer story with AI — keeping the human voice out front?”
(Let groups self-select to the role they care about — gives them a shared reason to talk + a soft reason to keep in touch, a stranger-friendly version of Isabelle’s same-PLC point.)
THE BOOSTS (the heart of it — print on the back, with room to add your own)
Two flavors in every area. AI boosts explore, they don’t shortcut — they make the AI ask, not answer. Human boosts force real interaction — they only work if you talk to someone.
| Area | 🧑 Human boost | 🤖 AI boost (paste into any chatbot) |
|---|---|---|
| Design | Borrow an outsider — someone from another table grills your problem with questions until it sharpens. | “Ask me 5 questions to find what I’m missing about this problem. Don’t suggest solutions yet.” |
| Build | Co-sketch — build it on paper with a partner, yes-and-ing each other. | “Give me 3 rough versions of this idea. Don’t tell me which is best — I’ll cut and remix.” |
| Advocate | Pitch to a face — deliver your 30-sec pitch to a real person; ask for a gut reaction + one hard question. | “Be a skeptical school-board member. Poke holes in my pitch and ask me the hard questions.” |
Spare boosts to swap in: Design 🧑 “silent gallery” (others post questions on your sheet, no talking); Build 🤖 “what breaks?” (AI lists 3 ways your prototype fails); Advocate 🧑 “steelman swap” (a neighbor argues your idea back to you).
ROVING FACILITATOR CARD (Reuben + Josh — pocket reference)
You’re not running tables; you’re nudging the choosing. Watch for:
- A group frozen on which boost? → “Pick one and go — you’ll feel the other either way.”
- One voice dominating? → hand the quiet two a human boost to run (it needs their hands).
- A group defaulting to AI for everything? → “Try the human boost on this one and feel the difference.” (The contrast is the lesson.)
- Noise drowning a table? → point them back to the card; it’s self-advancing by design.
- Drift: pull them to the reflection — which boosts did you pick, and why? The pattern of choices is the AI-Ready Graduate making informed decisions in real time.
Marketing the Experience → Conversion
The pavilion run is a taster. The 60-minute core delivers the “aha” — you made three real AI-or-human calls and saw your own pattern. The last 30 minutes turn that into a next step. The goal: people leave wanting the whole experience or wanting to run it at their school.
Two doors (make the pitch concrete at 1:00):
- “I want the whole thing.” A full 2-hour Flash Lab on a real topic, run for your staff or district. → inquire / book a session.
- “I want to run it myself.” The toolkit is open and free (designkit.stanford.edu); the train-the-trainer program gets you facilitator-ready. → take the kit, join TTT.
One line to say out loud: “What you just did in 60 minutes, you can run with your whole staff next week. The kit’s free, and we’ll train you.”
The draw (pulling people in from the aisle):
- “Human or AI?” board — at the aisle edge, name a real teaching task (“Grade a stack of essays” / “Plan tomorrow’s lesson” / “Comfort a struggling kid”) and let people drop a dot under 🧑 or 🤖. A 5-second commitment, instant social proof (see the split), and the hook: “Come find out where your gut was right — and where it wasn’t.” The choice mechanic that runs the workshop is also what pulls people in.
- The filling sticky board (“I chose AI for… / a human for…”) is live social proof — a crowd draws a crowd.
- A takeaway people carry out: a printed boost card (the human + AI boosts) + QR to the toolkit. A walking ad that also delivers the CTA.
Capture, or the leads leak: one QR → a 20-second form (name, email, school/role, “I want to: run it myself / bring it to my district / just stay in touch”). Feeds the Flash Lab partner CRM + the facilitator pipeline. You + Josh sort the people who lingered past the hour on the spot.
Positioning (what gets them excited): it’s not a demo of AI tools — it’s an experience that changes how you decide when to use AI, and you can re-run it yourself. Topic-agnostic, hands-on, free to take home.
This is the “Activation Stage” concept (action item 18f475ad) made concrete, feeding the train-the-trainer pipeline (GPS Goal 3).
PRINT / LOGISTICS
- 1 card per table (cardstock, double-sided: front = the 3 areas, back = the boost menu + the “run it yourself” CTA), + a sticky pad + a fat marker per table.
- A shared “When we chose AI vs. a human” board for the closing stickies — doubles as the takeaway artifact + a photo for the recap.
- Cap tables at 3 chairs. If overflow, add tables, don’t add chairs.
- Bring the printed cards from home — don’t rely on on-site printing.
Drafted 2026-06-18 (pre-OOO). Corrected to the real Flash Lab mechanic — human boosts vs. AI boosts, chosen per area (Design/Build/Advocate). Sized to a 90-min slot with the meat in the first 60 (people drift at the hour), plus a conversion layer (taster → full experience / take-it-to-your-school). Built off Isabelle’s DCI feedback (Jun 15–16 thread) + Josh’s note that paper scaffolds + roving nudges carried DCI.
prep/2026-06-18-iste-flashlab-table-card.md