Pre-built. Retail-specific.
One playbook per job.

1. Confirm the deviation against the per-store baseline (rule out promo pull-forward and weather lift).
2. Compute the recommended replenishment quantity using the cluster-level demand forecast.
3. Match against current open POs; suppress if a sufficient PO is already in transit.
4. Open a case; assign to the category buyer or auto-routed by lane.
5. Draft the replenishment order against the WMS / supplier portal.
6. Gate the write-back on buyer approval; default 4-hour SLA before escalation.
7. Verify on-shelf reads after delivery; close the case when fill rate returns above threshold.

Fill rate delta vs. baseline over the 14 days following close. Case is "outcome-positive" if fill rate ≥ pre-agreed threshold for ≥ 7 of those 14 days.

1. Pull the matching PO, GRN, and contract clause.
2. Compute the dispute amount; classify the cause (over-billing, short-ship, off-contract pricing, freight misallocation).
3. Assemble the evidence packet: invoice scan, GRN signature, contract clause excerpt, calculation table.
4. Draft the dispute letter using the vendor-specific template on file.
5. Open a case; assign to the AP analyst by vendor portfolio.
6. Gate the dispute send on AP approval; default 24-hour SLA.
7. Track the dispute through credit memo or settlement; close when funds recovered or written off.

Dollars recovered as credit memo or short-pay. Case is "outcome-positive" if recovery ≥ 80% of disputed amount within 60 days.

1. Decompose gross promo lift into incremental, cannibalized, pulled-forward, and pantry-loaded volume.
2. Identify the overlapping promo or adjacency that's eating the lift.
3. Model the scenarios: kill A, kill B, narrow A's scope, change A's discount.
4. Recommend the lowest-regret action; show the expected net lift under each.
5. Open a case; assign to the merchandising lead for the category.
6. Gate the cancellation on merchandising approval (writes are revenue-affecting).
7. Track post-cancellation lift; close when net category lift returns above threshold or promo ends.

Net category lift over the remainder of the promo window vs. the modeled counterfactual. Case is "outcome-positive" if cancellation moves net lift above threshold within 48 hours.

1. Cross-check sales drop against the compliance signal to confirm planogram cause (rule out promo, weather, traffic).
2. Render the corrected planogram for the affected category and store.
3. Open a case; assign to the district manager.
4. Push the corrected planogram to the field-ops app with a 24-hour reset SLA.
5. Notify the store via the standard ops channel.
6. Schedule the next-cycle photo audit; auto-trigger when due.
7. Close the case when photo audit confirms compliance and category sales recover above cluster benchmark for 7 days.

Category sales delta vs. cluster benchmark over the 14 days post-reset. Case is "outcome-positive" if sales gap closes by ≥ 75% within that window.

1. Decompose total shrink into theft, damage, spoilage, dock-receipt mismatch, and admin error.
2. Identify the dominant cause; isolate its most likely interval (day, daypart, dock vs. floor).
3. Pull the matching transactions, GRN events, and POS exceptions.
4. Pull the camera windows for the relevant intervals (where camera integration exists).
5. Open a case in the LP system; pre-fill the evidence chain.
6. Recommend the next investigation step (interview, dock audit, floor recount).
7. Track resolution through the LP system; close when corrective action is logged and shrink trends back toward baseline for 14 days.

Shrink rate trend over 30 days post-action vs. pre-action baseline. Case is "outcome-positive" if shrink moves toward estate baseline by ≥ 50% of the gap within 30 days.

1. Forecast end-of-season sell-through under the current ladder.
2. Model alternative ladders: shallower-but-earlier (5% at W5, 15% at W8, 30% at W12) vs. status quo.
3. Pick the ladder that maximizes margin-weighted sell-through.
4. Open a case; assign to the merchant for the category.
5. Gate the new ladder on merchant approval (writes are price-affecting).
6. Push the ladder into the pricing engine; sync POS price rules.
7. Track sell-through against the modeled curve; close when SKU clears at modeled or better margin.

Margin-weighted sell-through vs. the modeled counterfactual (status quo ladder). Case is "outcome-positive" if margin recovered ≥ modeled, less the cost of the earlier markdowns.