Shelly guideToday's plan
Findings are only useful if someone acts on them. Shelly's strategist turns model outputs into prioritised, costed and owned actions, the way a commercial manager would brief a store.
At a glance
- Priority order: compliance first, then P1 → P2 → P3, then largest $ impact
- Each action has a 'why' built from the evidence, a weekly $ impact where it can be estimated, and an owner
- Ticks are saved in your browser for this week only
- 'Ask why' explains any action in the chat
How it works
- Rules map each finding type to an action: recall → pull and quarantine; channel outage → check the tablet at shift start; stock-out → emergency order; shrinkage → re-count and CCTV; count gain → check GR/DDN; waste blow-out → cut order ~30% and move markdowns earlier.
- KPI breaches add actions too: waste over target, margin under target, budget gap below −3%.
- The reorder list adds one P1 when lines are below lead-time cover. Rostering and range review add P3 items.
- Actions that meet the governance policy become decisions that wait for your confirmation (see Decisions).
How to read it
- Start at the top and don't skip compliance items.
- A $ figure of '–' means the value can't be estimated honestly (e.g. a GS1 barcode fix). It doesn't mean the item is unimportant.
Limits & safeguards
- $ impacts are estimates from the week's data, not audited figures.
- The rules encode common convenience-store practice. Adjust them in strategist.py for a different format.
Ask Shelly
What should I do today?Why is the Uber Eats check a P1?
Type or say these in the chat on the main page.