--- stepsCompleted: - step-00-conversion - step-02-scope-decision-design - step-03-security-data-handling - step-04-data-readiness - step-05-model-v1 - step-06-feasibility-baseline - step-07-stability-sanity - step-08-robust-scenarios - step-09-exec-pack - step-10-execution-adoption - step-11-closeout - step-12-finalize - step-03-edit-scope-kpis - step-13-finalize - step-04-edit-data-handling - step-05-edit-data-readiness - step-07-edit-feasibility - step-09-edit-scenarios - step-10-edit-exec-pack - step-11-edit-execution - step-12-edit-closeout - data-swap-M2-M4-M5 - data-swap-faker-M2-M4-M5 lastStep: data-swap-faker-M2-M4-M5 date: '2026-01-14' user_name: Debian pilot_id: pharma-biotech-pilot-2026-01-14 status: FINAL runResult: success-closed-out --- # Pilot Blueprint Pack ## 0. Cover and Metadata - Shareable copy: yes (anonymized) - Freeze status: Final demo baseline (no further edits; data swap only) - Workspace initialized for conversion draft (no source materials). - Pilot ID: pharma-biotech-pilot-2026-01-14 - Customer/Site: Anonymized Site - Version: v0.1 - Date: 2026-01-14 - Owner(s): Founder; Product/Tech lead - Approver(s): Self - Module versions: constraint library v1; templates v1; workflow v1 ## 1. Executive Summary - Promise (1-2 sentences): Decision-grade planning that optimizes reliability under uncertainty with explicit tradeoffs. - Decision(s) enabled: Campaign sequence and staffing coverage for the next 6-10 weeks, with risk vs throughput vs overtime tradeoffs. - Top-line outcomes: Faster alignment, quantified bottlenecks, scenario tradeoffs tied to KPI table, and explainable risk drivers. **Disclaimer:** This pack provides decision support only. It makes no compliance claims, is not a GMP validation system, and does not replace MES/ERP systems. ## 2. Pilot Charter - Beachhead use case: Tactical campaign planning for a pharma/biotech site (6-10 week horizon). - Scope boundaries: Single product family; end-to-end process (DS -> DP -> Fill -> QC -> Release); top resources and constraints only. - Out of scope: Multi-site planning, real-time MES integration, GMP validation claims, or ERP/MES replacement. - Primary decision: Campaign sequence and staffing coverage for the next 6-10 weeks. - Secondary decisions: Identify whether cold storage slots or QC delays are the limiting lever; choose weekend coverage or shift capacity changes. - Decision owners and cadence: Supply Chain/Planning lead with Production and QA/QC; weekly planning review. ## 3. Stakeholders and Operating Cadence - Customer champion: Planning/Supply Chain lead (TBD) - Internal owners: Founder; Product/Tech lead - Cadence and sign-off points: Weekly working session; sign-offs at M1, M2, M4, M5, M6. ## 4. Success Metrics - Reliability metric (definition + formula): Probability of zero critical violations. Reliability = 1 - P(hold-time or deadline violation). - Performance metric (definition + formula): Throughput = total batches completed / week. - Human metric (definition + formula): Expected overtime hours per week = sum(max(0, required hours - scheduled hours)). - Baseline measurement method: Current Excel plan for the same horizon (fallback: last comparable period). - Definition of Done: M1-M7 approved; baseline accepted; 2-3 scenarios produced; recommendation tied to KPIs; adoption test defined. ## 5. Data Inputs and Assumptions - Data sources provided: Faker stand-in CSVs generated under inputs/real and loaded into inputs/raw (not real site data) - Completeness score/gaps: Faker stand-in data used; M2 reopened pending re-approval - Assumptions (list + approver): See assumptions register (stand-in data approved by self) - Data readiness decision: GO (stand-in data; pending re-approval) - Data handling plan: - Required data types: campaigns/orders, process steps, resources/capacities, calendars, critical constraints - Anonymization: mask site/product names - Storage location: local pilot workspace (anonymized path) - Access controls: founder + product/tech lead only - Retention/purge policy: delete or archive after pilot closeout - Security review: not run (optional) ## 6. Model Specification - Process model overview: DEGRADE v1 macro steps (DS -> DP -> Fill -> QC -> Release) with simple precedence chain. - Resource model: Top resources only, aggregated capacities (suite/room, fill line, cold storage slots, QC lab capacity, operators/QA/QC as pooled capacity). Single shared calendar assumed for now. - Constraint coverage (Hard/Soft/Monitor): - Hard: precedence, hold-time max, time windows, capacity conflicts, cold storage capacity, QC capacity - Soft: skills/qualification availability, changeover/cleaning time - Monitor: separation-of-duties, campaign grouping/batching ## 7. Uncertainty and Risk Policy - Uncertainty representation (min/typical/max): - Risk metric choice: - Risk stance (threshold/lexicographic): ## 8. Planning Outputs and Scenarios - Deterministic baseline schedule: exports/baseline-schedule.csv (faker stand-in data) - Robust scenarios (2-3 minimum): conservative, balanced, aggressive (faker stand-in data) - Risk decomposition (top drivers): attachments/risk-decomposition.md (faker stand-in data) - Stability sanity check: attachments/stability-sanity-check.md (synthetic demo) ### Scenario Comparison (KPI Summary) | Scenario | Reliability | Performance | Human Load | Notes | | --- | --- | --- | --- | --- | | Conservative | 0.90 | 8 batches/week | Low overtime | Synthetic demo | | Balanced | 0.80 | 10 batches/week | Moderate overtime | Synthetic demo | | Aggressive | 0.65 | 12 batches/week | High overtime | Synthetic demo | ## 9. Recommendation - Recommended scenario: Balanced (P_success 0.80) based on KPI tradeoff and risk decomposition - Tradeoffs (what improves, what risk is accepted): +throughput vs moderate overtime and accepted risk; conservative reduces risk at cost of throughput - Monday-morning next steps: Replace synthetic inputs with site data, re-run baseline + scenarios, re-approve M2/M4/M5, then execute adoption test ## 10. Pilot Execution Plan - Re-plan cadence: Weekly planning cycle; mid-week review if deviations occur. - RACI: Data updates (planner), model updates (product/tech lead), scenario generation (product/tech lead), decision approval (planning lead + QA/QC), exec readout (founder). - Measurement and reporting: Compare against current Excel baseline; report weekly deltas on KPIs. - Adoption test evidence: attachments/adoption-test-evidence.md (explicit pass criteria) ## 11. Validation Checklist and Sign-offs - M0: PASS (Owner/Approver: self; 2026-01-14) - M1: PASS (Owner/Approver: self; 2026-01-14) - M1.5 (if used): PASS (Owner/Approver: self; 2026-01-14) - M2: PASS (Owner/Approver: self; 2026-01-14) - M3: PASS (Owner/Approver: self; 2026-01-14) - M4: PASS (Owner/Approver: self; 2026-01-14) - M4.5 (if used): PASS (DEGRADE placeholder; Owner/Approver: self; 2026-01-14) - M5: PASS (Owner/Approver: self; 2026-01-14) - M6: PASS (Owner/Approver: self; 2026-01-14) - M7: PASS (Owner/Approver: self; 2026-01-14) - M8 (if used): PASS (placeholder closeout; Owner/Approver: self; 2026-01-14) - Approver log: M1 approved by self on 2026-01-14; M1.5 approved by self on 2026-01-14; M2 (DEGRADE) approved by self on 2026-01-14; M3 approved by self on 2026-01-14; M4 approved by self on 2026-01-14; M4.5 approved by self on 2026-01-14; M5 approved by self on 2026-01-14; M6 approved by self on 2026-01-14; M7 approved by self on 2026-01-14; M8 approved by self on 2026-01-14 ## 12. Appendices - Data templates: - KPI glossary + formula registry: - Constraint library references: - Optional add-ons: - SOP extraction notes - Integration/security notes: data handling plan approved; no external references used - Capacity/investment scenarios - Changeover/sequence-dependent modeling - As-executed comparison - ROI calculator sheet - Closeout artifacts: attachments/pilot-run-report.md, attachments/library-merge-log.md (synthetic closeout)