Sample walkthrough: resolve a capacity conflict

This walkthrough uses only the fictional Northstar plan. It demonstrates the decision the sample supports; it does not forecast delivery or recommend a staffing action.

Start with the seeded constraint

Start from the seeded plan by choosing Reset sample, then Reset sample data, if the browser contains earlier local changes; the confirmation replaces the saved local scenario with the fictional fixtures. Open Plan and find Noah Williams in the Sep 28 column. His weekly capacity is 5 person-days, time off is 5, and non-project commitments are 0. The result is 0 available. The same cell still includes 1 day of Signal accounts work, so the planning surface marks the person-week as a conflict and Signal accounts as at risk.

The useful question is not “can the dashboard make the work fit?” It is “what choice should the planning group inspect before making a commitment?” The zero availability is a capacity signal. It does not predict a missed date or identify a person who should absorb the work.

Make an explicit draft choice

Change Noah Williams, Signal accounts, Sep 28 person-days from 1 to 0. Allocation fields accept 0–10 person-days in half-day increments, so the sample supports a precise change without inventing a recommendation. Zero removes that allocation entry; the field appears empty with a “0” placeholder. The draft removes the assigned day, but Noah still has zero available days because time off is unchanged. The initiative-risk summary recalculates.

This is a tradeoff: removing the day resolves Noah's Sep 28 conflict in this scenario, but it does not establish where the work goes or whether the initiative should retain its scope. Signal accounts can remain at risk because its other allocations include a separate overloaded person-week. A planning lead can use the draft to compare options, then discuss deferral, a different week, or a different allocation with the people who own the work.

Compare, apply, or discard

Open Scenarios to compare the editable draft with the applied baseline, then return to Plan to choose Apply draft or Undo draft. Until Apply draft is selected, the original plan remains the baseline. Choose Apply only when the group intends to retain this fictional scenario in the current browser; applied plans persist locally and any prior “undo reset” recovery message is cleared because the new scenario is now the baseline. Choose Undo when the option was only an inspection, or use Export draft to create a local JSON discussion copy. For the one-cell Noah edit, the export reports one allocation-cell change, zero availability-cell changes, and the boundary Fictional working draft; local export only. The interface announces that one draft change was exported.

The distinction is deliberate. Drafts are not retained across refreshes until they are applied, and exported data never leaves the device through the sample. Editing an allocation or availability constraint after export clears the old export notice, so the visible status never claims that an earlier file represents the current draft. Saved plans also require every visible week key and case-insensitive unique initiative codes; malformed payloads fall back to the sample with a warning. The JSON is a full planning inspection copy rather than a minimal decision summary, so review its contents before sharing it outside the walkthrough. A resolved conflict is still not a delivery forecast. The next product question is whether planning practitioners can explain that boundary and the arithmetic without help; the evaluation plan defines the proposed study and guardrails. If the saved payload is malformed, Apply remains blocked until the user explicitly resets the sample, preserving the recovery boundary. A browser storage access failure is handled separately: the current tab remains available for inspection, but later persistence may be unavailable.