Executive Summary
Analysis window: 2025-12-18 → 2026-06-16 (180 days). Zero PR outcomes — all ~310 workflow-created PRs in the sample lack Closes #N links. Among 50 sampled safe-output report issues, 46 closed outcomes delivered 2,355 value units at 3,699 AIC (partial data; 70% of runs missing AIC). Observability (OTLP Validation) is the highest-efficiency category at 3.75 value-units/AIC. Workflow-optimization (Ambient Context Optimizer) consumed the most AIC (2,625 AIC, 71% of total) with the lowest return (0.32 value-units/AIC), driven by three anomalously expensive runs that have since been resolved.
Summary
| Metric |
PRs |
Safe-output Issues |
Combined |
| Analyzed (linked issue / in scope) |
0 |
50 |
50 |
| PRs excluded — no linked closing issue |
310+ |
— |
310+ |
| Skipped (open / pending) |
— |
4 |
4 |
| Mapped to objectives |
0 |
41 |
41 |
| Unmapped |
0 |
5 |
5 |
| Accepted outcome count |
0 |
46 |
46 |
| Total outcome value |
0 |
2,355 |
2,355 |
| AI Credits (partial) |
0 AIC |
3,699 AIC |
3,699 AIC |
| Impact Efficiency |
— |
0.637 |
0.637 value-units/AIC ⚠️ partial |
⚠️ 32 of 46 closed runs had no AIC in footers. Impact Efficiency is a lower-bound estimate (true IE is lower).
What We Worked On
🔭 observability — value 70/outcome — Best IE
- 19 accepted | 1,330 total value | 354.7 AIC | IE = 3.750
- #37755, #37066, #36850 (OTLP Validation & Portfolio Yield reports)
⚙️ workflow-optimization — value 50/outcome — Most AIC
- 17 accepted | 850 total value | 2,624.5 AIC | IE = 0.324
- #39279, #39142, #38926 (Ambient Context Optimizer)
- ⚠️ Three runs in 2026-06-05–07 consumed 723–937 AIC each (vs. ~20 AIC steady-state). Cost since resolved.
🤖 automation — value 35/outcome
- 5 accepted | 175 total value | 283.0 AIC | IE = 0.618
- #37700, #37502, #37089 (PR Triage, Health Dashboard, aw-compat)
Top 15 outcomes by outcome value
Unmapped outcomes
Interpretation
Accepted outcome count (46) signals a productive system across three mapped objective categories. However it misses the entire PR bucket (zero qualifying PRs) and covers only ~50 of potentially hundreds of safe-output issues.
Impact Efficiency (0.637 value-units/AIC) is more informative: it correctly identifies that observability workflows deliver 11× better return per AIC than workflow-optimization. The metric is unreliable at full fidelity because 70% of runs lack AIC data, but it unambiguously flags the Ambient Context Optimizer cost spikes as the dominant AIC sink.
Highest value per AIC: observability (OTLP Validation) at 3.750 — low cost, high signal value.
Most AIC, weakest return: workflow-optimization at 0.324 — historical spikes dominate the denominator.
Data quality
PR closing-issue links: Critical gap. 310+ workflow-created PRs sampled; zero had Closes #N body keywords. All PR outcomes excluded. Adding issue-closing links to Copilot PR templates would unlock this outcome bucket.
Safe-output label mapping: 5 of 50 issues (10%) unmapped. daily-status, workflow-health, outcomes, cookie labels absent from .github/objective-mapping.json.
AI Credits: Live fallback (no precomputed gh aw logs --json data found). AIC sourced from issue-footer text. 32 of 46 runs (70%) missing AIC → denominator 3,699 AIC is a lower bound; true Impact Efficiency is lower than 0.637.
Analysis window: 2025-12-18 → 2026-06-16 · Sample: 50 report-labelled issues + 400 workflow PRs
AIC source: issue-footer live retrieval (fallback; deterministic precomputed data unavailable)
Generated by 📊 Impact Efficiency Report run #5 · 2026-06-16
Generated by 📊 Impact Efficiency Report · 1.1K AIC · ⌖ 26.4 AIC · ⊞ 29.5K · ◷
Executive Summary
Analysis window: 2025-12-18 → 2026-06-16 (180 days). Zero PR outcomes — all ~310 workflow-created PRs in the sample lack
Closes #Nlinks. Among 50 sampled safe-output report issues, 46 closed outcomes delivered 2,355 value units at 3,699 AIC (partial data; 70% of runs missing AIC). Observability (OTLP Validation) is the highest-efficiency category at 3.75 value-units/AIC. Workflow-optimization (Ambient Context Optimizer) consumed the most AIC (2,625 AIC, 71% of total) with the lowest return (0.32 value-units/AIC), driven by three anomalously expensive runs that have since been resolved.Summary
What We Worked On
🔭
observability— value 70/outcome — Best IE⚙️
workflow-optimization— value 50/outcome — Most AIC🤖
automation— value 35/outcomeTop 15 outcomes by outcome value
observabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityobservabilityUnmapped outcomes
daily-status,reportnot in mappingreport,outcomesnot in mappingreport,outcomesnot in mappingworkflow-health,cookienot in mappingworkflow-health,cookienot in mappingInterpretation
Accepted outcome count (46) signals a productive system across three mapped objective categories. However it misses the entire PR bucket (zero qualifying PRs) and covers only ~50 of potentially hundreds of safe-output issues.
Impact Efficiency (0.637 value-units/AIC) is more informative: it correctly identifies that
observabilityworkflows deliver 11× better return per AIC thanworkflow-optimization. The metric is unreliable at full fidelity because 70% of runs lack AIC data, but it unambiguously flags the Ambient Context Optimizer cost spikes as the dominant AIC sink.Highest value per AIC:
observability(OTLP Validation) at 3.750 — low cost, high signal value.Most AIC, weakest return:
workflow-optimizationat 0.324 — historical spikes dominate the denominator.Data quality
PR closing-issue links: Critical gap. 310+ workflow-created PRs sampled; zero had
Closes #Nbody keywords. All PR outcomes excluded. Adding issue-closing links to Copilot PR templates would unlock this outcome bucket.Safe-output label mapping: 5 of 50 issues (10%) unmapped.
daily-status,workflow-health,outcomes,cookielabels absent from.github/objective-mapping.json.AI Credits: Live fallback (no precomputed
gh aw logs --jsondata found). AIC sourced from issue-footer text. 32 of 46 runs (70%) missing AIC → denominator 3,699 AIC is a lower bound; true Impact Efficiency is lower than 0.637.