BMAD-METHOD/bmad/bmm/agents/hand-off/DEMO-INSTRUCTIONS.md

62 lines
3.4 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Demo Instructions — Capture → Transcribe → Convert → Publish (with fallback)
Purpose: step-by-step script for a live demo to PM or stakeholders. Use a test LinkedIn account for direct publish testing.
Pre-demo checklist
- Ensure serverless endpoints are deployed and env vars set (OPENAI_API_KEY, WHISPER_API_KEY, LINKEDIN_CLIENT_ID/SECRET, STORAGE creds).
- If LinkedIn direct publish will be shown, have a test LinkedIn account and ensure the server app is registered with correct redirect URI.
- Mobile device with Expo Go or installed dev build.
- Clipboard app/permissions allowed on device.
Demo script (1015 minutes)
1. Quick intro (30s) — One sentence: app helps professionals capture daily notes (text + voice) and converts them into LinkedIn-ready posts with privacy-first defaults.
2. Capture flow (12 min)
- Open app Home. Create a quick text entry (or press record for voice).
- If recording: press record → speak for 20s → stop → press Transcribe.
- Show local-first behavior: entry saved locally; no external processing until user confirms.
3. Transcription (1 min)
- Transcribe job submits to server via signed-upload and Whisper. Show job status spinner.
- When transcript returns, show editable transcript screen and anonymize toggle. Demonstrate redaction.
4. Convert to post (2 min)
- Tap "Convert to LinkedIn post" → server proxies to OpenAI with system prompt. Show generated variants (Variant 1 & 2), suggested hashtags, and CTA.
- Edit a variant inline (minor tweak) in draft editor.
5. Publish (2 min)
- If LinkedIn connected: press Publish → server posts via LinkedIn UGC API → confirm published and show post URL or success toast.
- If LinkedIn not connected: use fallback: Copy to clipboard → Open LinkedIn app (or show share sheet) → paste & post manually.
6. Analytics & logs (1 min)
- Open analytics dashboard: show counts for entries, conversions, and publish events.
- Show processing logs in settings (transcribe/generate/publish events with timestamps & consent).
Demo wrap-up (30s)
- Remind PM of decisions needed: retention default, LinkedIn publish default opt-in vs auto, and monetization approach.
- Offer to share sprint tickets and server API spec for execution.
Troubleshooting tips
- If Whisper/transcribe fails: show job status and retry option; confirm signed upload worked and audio accessible in storage.
- If OpenAI generation fails: show friendly error and option to retry with lower variant count.
- If LinkedIn publish fails: show fallback "Copy & Open LinkedIn" flow.
Sample entries to use during demo
- "Today I resolved a recurring onboarding issue that caused users to drop off on day 1. Turned out the default timezone handling was silently corrupting event timestamps. I shipped a quick fix and added a telemetry check — early data shows a 12% improvement in day-1 retention. Learned: small telemetry + quick rollback plans beat speculative rewrites."
- "I experimented with 5-minute daily standups. The team trimmed ~3 hours of weekly meetings and improved follow-through."
- "I published a vulnerable first post and received great DMs that changed my perspective. Authenticity wins."
Post-demo artifacts to hand the PM
- `SPRINT-TICKETS.md` (import to tracker)
- `SERVERLESS-API-SPEC.md` (developer reference)
- `PROMPT-TEMPLATES.md` (prompt & model guidance)
---
Use this file as the demo script during the PM meeting.