PROGRAM · BP-001
AI 101 · the lunch hour.
A 60-minute lunch-and-learn built on judgement, safety, and craft.
The hour, minute-by-minute.
-
00:00 → 08:00
Movement I · How LLMs actually work.
A jargon-light account of next-token prediction, context windows, and the places models quietly fail. No anthropomorphising. No magic. By minute eight, the room knows why hallucinations happen.
Facilitator note — use the spreadsheet-LLM demo for technical audiences; use the "guess the next word" exercise for non-technical rooms.
-
08:00 → 18:00
Movement II · Which model, when.
ChatGPT, Claude, Gemini, Copilot — how to choose the right tool for the task, and what each model is genuinely better at. Includes a live "race the three" prompt against a real task pulled from the pre-class survey.
Facilitator note — verify the survey covers tooling currently licensed; do NOT recommend tools the team can’t actually access.
-
18:00 → 32:00
Movement III · The four-ingredient prompt framework.
Role, task, constraints, output format. Demoed live; built on the spot against a survey-supplied task. We rewrite a generic prompt and a structured one side-by-side and watch the difference. Each attendee leaves with the template in their notebook.
Facilitator note — this is the section attendees quote in the post-session survey. Anchor on it.
-
32:00 → 47:00
Movement IV · Role-relevant use cases.
Three live worked examples pulled from the pre-class survey. The room watches AI applied to tasks they actually do — not generic demos. Every example ends with the room voting whether to send, edit, or scrap the AI output.
Facilitator note — use the room’s vote to teach judgement, not the model output. The vote is the lesson.
-
47:00 → 60:00
Movement V · Data, risk, one redesign.
A grounded conversation on data hygiene and risk (no theatre, no fear). Then every person writes down one workflow they have committed to redesign by Friday. Names go on a shared list. We send a one-question follow-up survey on day seven.
Facilitator note — if the room balks at the “by Friday” commitment, swap to “by next session” — never to “eventually.”
Pre-class survey (3 min, per person).
- What’s your role — and what does “a busy week” look like?
- Which AI tool(s) do you use today, if any?
- Your confidence using AI, 1 – 5.
- One specific task you’d like help with on Monday.
- One thing you’re worried about (data, risk, replacement — anything).
Responses shape the live demo. Deleted within 30 days of delivery. Never pasted into public AI tools.
What each person leaves with.
- The four-ingredient prompt framework, written down.
- Three role-relevant prompts they saw demoed.
- One workflow they’ve committed to redesign.
- A clear sense of which AI tool to use, and when.
- Permission to ask “is this a good place for AI?” out loud.
- Honest awareness of where the models still fail.
Guarantee. If under 70% of attendees say the session saved them at least one hour in the following week, you don’t pay. Measured by a single-question email on day seven.
Data. Survey responses are used only to shape your session and are deleted within 30 days of delivery. We do not paste your team’s details into public AI tools.