# Trainer guide
Use syllabus.md and complete setup-checklist.md before delivery. These materials do not provide configured cloud resources or a complete backend baseline.
Session 1 is four hours for non-technical business users. Start with everyday event sources, then use the fictional business pack. Teach GenAI/LLMs, conceptual RAG and the need to check citations. Avoid developer terminology. Core project: three verified findings, action list, open question and labelled recommendation. Six teams of 3–5 fit 20–30 participants. Six five-minute presentation slots run 12:25–12:55.
Session 2 is eight elapsed hours for developers. Keep one event app, teach AI Studio first, then use a static export with Firebase Hosting and a callable Cloud Function. Rehearse the supported CLI/runtime, deployed secret, Auth, rules and client/function region. A generated server app cannot simply be copied into static Hosting; provide a compatible baseline. Cloud Run is an alternative only.
Demonstrate one operation, let learners repeat, inspect the result, then offer a retry. Explain each feature's purpose before clicking. Keep project time for finishing and testing existing work. Five eight-minute presentation slots run 16:15–16:55.
Assessment both sessions: correctness/evidence 35%; useful required outputs 30%; verification/limitations 20%; explanation/presentation 15%. Session 2 uses five core checks; log failed/not-run results honestly. Pair blocked learners and record help received.
Animations in core slides are planned storyboards. Existing technical animations remain in the optional extension deck. Use live demonstration while new animations are pending.

Session 1: business cases now 10:50–11:10 (demonstrate one; learners choose one mini exercise); reuse 11:10–11:20 (prepare an output in advance if needed); best practice 11:20–11:35 (8 minutes explanation/examples, 7 minutes practice). Preserve the 11:35 project and 12:25 presentations. Distinguish RACI ownership from Bloom’s learning-objective taxonomy adapted to prompting. Require one appropriate framework within the existing kit, not extra reports.

Emphasise the framework benefit: learners can name a familiar structure or thinking task and describe the outcome in one sentence. Start with the short examples; demonstrate one follow-up only if needed. Avoid implying that a framework guarantees accuracy or removes the need for context.

Module structure: each teaching module now opens with observable learning objectives and closes with a takeaway, a “Show you can” evidence check and a transition. Use these within the existing lesson/practice times, not as extra activities. If the check is incomplete, give feedback and let the learner retry or record help needed. Animation placeholders illustrate a process; they do not replace this teaching structure. Breaks and the resource list do not need separate objectives.