A company wants to make better decisions by pooling its employees' private estimates of uncertain things — how long a project will really take, how likely a launch is to hit its date, what a cost will come in at. The obvious approach is to ask people and average the answers. But employees are not neutral: some own the projects they are estimating, some fear that an honest pessimistic number will be held against them or their team, and some will say whatever they think leadership wants to hear. Leadership can decide what is asked, how answers are collected, what (if anything) is rewarded or recorded, who sees what, and how the aggregate is computed and used.

Decision required: Design how to collect and combine these estimates so the aggregate is actually informative.

Success criteria: The aggregate tracks reality better than naive polling would, AND people keep participating honestly over time. Failure = the numbers are gamed toward what people want to report, OR people stop giving real answers because reporting feels risky.