# Google AI workshops — approved scope
## Session 1: AI-Powered Knowledge & Productivity
Focus: NotebookLM / Gemini Notebook and GenAI for business value.
Audience: business analysts, product managers, knowledge workers and non-technical stakeholders. 20–30 participants. No coding.
Duration: half day, four hours, 09:00–13:00 including a 15-minute break.
09:00 purpose/baseline; 09:15 GenAI, LLMs, grounding and conceptual RAG; 09:35 notebooks/sources/navigation; 10:05 prompting/citations/notes; 10:35 break; 10:50 contract, meeting and market examples; 11:10 outputs/reuse; 11:20 best practice with RACI and Bloom’s; 11:35 project; 12:25 presentations; 12:55 exit.
Project: Business Briefing Kit — three verified findings, action list, one open question and a labelled recommendation using one appropriate framework. Six teams present for four minutes plus one minute feedback.
## Session 2: GenAI App Development & Backend Integration
Focus: Google AI Studio + Firebase. Audience: front-end, back-end and full-stack developers. 15–20 participants; basic JavaScript and terminal skills assumed, no previous AI Studio experience required.
Duration: one day, eight elapsed hours, 09:00–17:00 including lunch and breaks (6.5 hours of sessions).
09:00 AI Studio basics; 09:30 system instructions/few-shot/settings/API keys; 10:30 break; 10:45 Build/static export; 12:00 Firebase setup/Auth/Firestore; 13:00 lunch; 14:00 callable Cloud Function + Gemini; 14:25 Hosting deployment/reliability; 15:00 break; 15:15 acceptance brief; 15:25 project completion; 16:15 presentations; 16:55 exit.
Project: one AI-enabled Connected Event App. Main route: Firebase Hosting frontend + callable Cloud Function + Gemini, with Auth and owner-only Firestore records. Five acceptance checks. Five teams present for five minutes plus three minutes Q&A.
Cloud Run is an alternative architecture discussion, not a second required deployment. Prompt refinement is included; model fine-tuning and production scaling implementation are beyond the day.
Both sessions use explain → example → practice → check → feedback. Trainer preflight: setup-checklist.md.