The platform follows the decision, not the report.
Why it matters when the plan changes
Organisational work usually arrives as a deliverable: a report, a deck, a review. The deliverable ends and the decision it was meant to inform carries on without it. Making the decision the unit rather than the document is what keeps the evidence, the choice and the result attached to each other long enough to learn anything. That record only produces learning if it survives past the moment the report was delivered.
The tension is that a case is harder to sell than a report. A report has a delivery date and a clear end. A case stays open until the outcome is known, which is uncomfortable commercially and is the only arrangement in which the next judgement can be better than this one. Philip E. Tetlock and the Good Judgment Project found that judgements recorded and later compared with what happened improve, while judgements that end at the deliverable do not; the case is what makes that comparison possible.
In practice
A division leader has to decide whether to combine two teams. The case records the question, the roles and dependencies in scope, what the evidence showed on both sides, the recommendation and its confidence, what was decided and by whom. Two quarters later the outcome is recorded against it, and the next similar decision starts from something real.
Evidence
Judgements that are recorded and later compared with what happened improve; judgements that end at the deliverable do not.
Philip E. Tetlock and the Good Judgment Project (2026)
What it cannot tell you
A decision case preserves what was known, recommended and decided, and what followed. It cannot establish that the outcome was caused by the decision rather than by conditions that changed anyway, and it does not judge whether the leader chose well. It records the sequence; interpreting it remains a human task.
How Atlas reads it
A case has two roles in it: the accountable leader who defines the question, reviews the evidence and decides, and the participant who contributes evidence and receives what is relevant to their own work. Match Reports, Team Reports and Collaboration Playbooks are views of the evidence held inside a case, which is why a separately ordered report does not require a separate product.
Questions
A decision case opens with one decision owned by a named leader, akin to the multi-year design behind Philip E. Tetlock's Good Judgment Project (2026), which tracked judgements against outcomes over time rather than closing at a single point. The scope follows the decision, not the chart.
Access follows the purpose agreed and the permissions set, not seniority. The accountable leader sees the decision view and its evidence; a participant receives what is relevant to their own work. A participant does not become visible to the organisation beyond what the permission model allows.
A report ends when it is delivered. A case stays open until the outcome is recorded, which lets the earlier judgement be compared with what happened. Philip E. Tetlock and the Good Judgment Project (2026) found that recorded judgements compared against outcomes improve, while judgements ending at delivery do not.
No. The case records what was permitted and available for this question, and marks what was not. Where signal is thin the view is marked blind rather than filled in, and the case states what evidence would resolve it and who would need to provide it.
The earlier judgement is preserved and a new one is made against the changed plan. Keeping both is the point: the comparison between them shows what the change actually did, which is information that replacing the earlier view would have destroyed.