Episode 1 — Retrospective Ritual
Purpose
- Capture lessons learned, strengthen standards, and explicitly shape the onramp for Episode 2.
Standard Work (Retrospective) — with Rationale
- Review objectives vs outcomes
- Rationale: Align expectations with results; calibrate scope.
- Identify what worked and why
- Rationale: Preserve winning patterns; scale them.
- Identify what didn’t and why
- Rationale: Target the next minimal refinements.
- Confirm standards deltas
- Rationale: Bake improvements into reusable artifacts.
- Define Episode 2 focus hypothesis
- Rationale: Maintain momentum and clarity of purpose.
Objectives vs Outcomes
- Objectives met: MVP kit, strict runbooks, citations/annex discipline, three dry runs.
- Underutilized potential: fuller mining of example offers to pre-fill matrix and decisions; richer CF04 assembly in earlier runs.
What worked (keep)
- Strict grammar asserts improved quality gates without changing flow.
- Standards Toolkit integration reduced context switching; templates enforced consistency.
- Annex assembly (attachments, quality docs, references) now explicit in CF04.
What didn’t (improve)
- Prefill before RFI: ensure all explicit offer/datasheet claims are harvested in CF2/CF3.
- MR alignment: when MR missing, provisional reconciliation must still be produced and flagged MR‑pending.
- CPQ mapping: produce provisional list from offers even with some unknowns, tagging RFI‑pending lines.
Standards deltas (applied)
- Runbooks and sequencing updated with asserts for: profile_set; reconciliation present; extended field presence; hazards matrix + weakest‑link; CPQ list; annex manifests; Appendix 2/3 autopopulation.
Episode 2 Focus (hypothesis)
- Shift from “demo correctness” to “demo completeness & saliency”:
- Increase prefill coverage and reduce RFI scope by deeper mining of offers/datasheets.
- Produce near client‑ready packet for a selected example, aligning to the final proposal standard.
Exit Criteria for Episode 2 (draft)
- Prefill coverage ≥80% of enforceable and component claims; RFIs only for genuine unknowns.
- End‑state packet includes complete annexes (Appendix 2/3; quality docs; reference list) and a populated executive summary.
- Metrics show adherence to asserts with zero critical misses.