Evidence-led executive learning · reviewed August 2026
Annotated 30-day experiment charter
Fictional proposal-writing experiment. Planning the run is the assessed artefact; these are illustrative assumptions, not observed results.
| Field | Exemplar | Why it works |
|---|---|---|
| Decision | Whether to extend the approved workflow to one additional trained proposal team. | Bounded authority, not enterprise scale. |
| Hypothesis | Source-linked drafting plus rubric review will reduce median total handling time without a critical unsupported service claim. | Connects intervention, outcome and quality boundary. |
| Design | 40 baseline and 40 run proposals from three approved service families; eight specialists across two teams; exclude confidential attachments and regulated advice. | Population, comparator and exclusions are explicit. |
| Measures | Draft-plus-review time, rubric quality, critical source errors, eligible-use adoption, overrides, boundary events and usable hours redirected. | Balances value, quality, adoption and risk. |
| Gate | Scale one team only if all thresholds pass; restrict and retest a failing family; stop for uncontained confidentiality or persistent critical errors. | Decision rules exist before results. |
Weak counterexample
“Give everyone access for 30 days and survey whether they liked it.” There is no decision, baseline, eligible work, quality evidence, boundary, stop authority or value conversion.