AI-PDLC Afternoon - Facilitator Notes (9 Oct 2026, 13:30-17:00)
Not for attendees.
What runs when
| Time | Block | Material |
|---|---|---|
| 13:30-14:15 | Content 45 min: welcome, why AI-PDLC, how it works (RICE, Working Backwards, L0-L3), Handbook + Navigator | Deck slides 1-22 (5-8, 12, 16-17, 21-22, 30 come from the AI-PDLC First Call Deck; 15 redrawn from it) |
| 14:15 | Lab phases 1-3: Voice of Customer, Read the Signals, Ideate (60) | Workshop Studio guide (QR slides 2 and 23), slides 23-24 |
| 15:15 | Break (15) | |
| 15:30 | Lab phases 4-5: Solution Design, Prototype (40) | Workshop Studio guide, slide 25 |
| 16:10 | Part 6 live: 6-agent pipeline from aws-samples/sample-AI-PDLC-using-AmazonQuick (15) | Slides 26-27 (flow overview slide 11), AI-PDLC-Quick-Agents.zip, KT demo workspace |
| 16:25 | Navigator tracks A/B/C (15) | ai-pdlc-navigator-skill.zip, slide 28 |
| 16:40 | Share back, pilot, survey (20) | pulse.amazon/survey/QIDUNJDM |
Workshop link: https://catalog.us-east-1.prod.workshops.aws/workshops/cd263623-2478-48f8-b5bf-f7cd56d731f6/en-US
Roles
- KT: facilitator, timekeeper, Part 6 live agent pipeline + Navigator demo
- Patis: floor support (sign-in, Quick Desktop install, datasets in the space)
- Narongrit: floor support (Parts 2-5, Working Backwards Skill install, Navigator tracks)
Checks at each table
- Part 1: space has all 3 datasets.
- Part 2: competitor report citations open; no uncited claim kept.
- Part 3: RICE scores read per factor, not only the total; personas labelled synthetic.
- Part 4: Quick Desktop signed in, Working Backwards Skill installed; headline rewritten by a human.
- Navigator: Track A uses the Part 4 PRFAQ; gaps stay as
_TBD_, not invented.
Before the day
- Run the full Workshop Studio lab once on a Free plan account (check agent-hour limits).
- Install the 6 aws-samples agents in the demo workspace (exact names: AI-PDLC Orchestrator, Discovery, Prioritize, Working Backwards, Spec Generator, Designer; Orchestrator last; instructions 11.7k-21.4k chars) and a Space; rehearse "Run the full AI-PDLC pipeline" on Fast Track with the Part 3 pain point; keep repo outputs/ (ClinicalFlow example) open as fallback.
- Install ai-pdlc-navigator-skill.zip in Quick Desktop and test the 3 tracks (480 files; skill was written for Claude Code).
- Decide how attendees get the Navigator zip (USB, share link, or the GitHub repo).
Backup (if Workshop Studio is unavailable)
AI-PDLC-Lab-Pack.zip (aws-samples Quick agents,
AnyCompany Mart data). Answer key in facts.json:
delivery/ETA 31 feedback + 26 tickets is the top pain point.