Use the action grid and filters
The "Products that need your attention" action grid is the primary working surface in AI Inventory Manager. It surfaces AI-prioritized replenishment recommendations so you can review, filter, bulk-select, and act on stock risks without manually scanning every product in your catalog.
Audience: Inventory Coordinators and center managers who review and act on daily replenishment recommendations.
Before you begin
Permission: Add or Edit (on the Inventory security object) β required to view and act on the action grid.
- AI Inventory Manager must be active for your center. If you do not see the module in your navigation, contact your administrator.
- Your organization must have completed the onboarding wizard and run at least one nightly refresh so that recommendations are available.
- You need Add or Edit access on the Inventory security object to interact with the action grid and create purchase orders.
- The action grid reflects data from the most recent nightly refresh. Changes to settings or product data take effect at the next nightly refresh, not immediately.
Steps
Open the action grid
- In the left navigation, select Inventory Dashboard.
- Select the Inventory overview tab.
- Ensure you are in center view β use the center selector at the top of the page to choose a specific center. The action grid is not available at organization or zone view.
- Scroll past the KPI tiles to the Products that need your attention section.
The grid loads with rows already prioritized: stock-integrity issues appear first, followed by Critical-status products ordered by ascending days of stock cover, followed by Low-status products.
Read the action grid columns
Each row in the action grid displays the following information:
| Column | What it shows |
|---|---|
| Product name | Product name and SKU |
| Type | Product type β Retail or Consumable |
| Category / Subcategory | The product's category and sub-category |
| Daily usage | Predicted daily usage, marked with an AI badge to indicate the value comes from the ML forecast |
| Stock / Reorder level | Current on-hand stock and reorder level, with a proximity bar showing how close stock is to the reorder level |
| Incoming stock | Quantity on open purchase orders not yet received |
| Days of stock left | How many days of stock cover remain at the predicted daily usage rate |
| Status | A status badge β Critical, Low, or Normal |
| Lead time | The vendor lead time used in the reorder calculation (days) |
| Suggested qty | Recommended order quantity, marked with an AI badge |
| Vendor price | Unit price from the selected vendor |
| Value to order | Suggested qty Γ vendor price |
| Confidence | A confidence badge showing the forecast reliability tier (T1 β High through T4 β Very Low) |
Note: The Suggested qty and Days of stock left values are calculated using deterministic reorder math applied on top of the AI forecast β they are not themselves AI-generated. Only Daily usage and the Confidence tier come directly from the ML model.
Filter the action grid
Use the filter bar above the grid to narrow the rows displayed.
- Select Categories to filter by product category or brand. You can also select Brands or Vendor from the same filter area.
- Select Product type to show only Retail, only Consumable, or Both.
- Select Stock status to filter by Critical, Low, or Normal.
- Select Confidence to filter by confidence tier:
- T1 β High β forecast is stable and reliable
- T2 β Medium β forecast is moderately reliable
- T3 β Low β forecast has higher variance; review before ordering
- T4 β Very Low β limited forecast data; treat the suggestion as a starting point only
- To search by name or SKU, type in the search field next to the filters. The grid filters in real time as you type.
Each filter you apply appears as a removable chip below the filter bar. To remove a single filter, select the Γ on its chip. To clear every applied filter at once, select Reset all.
Tip: Combining Stock status: Critical with Confidence: T1 β High is a reliable way to identify the highest-priority, highest-confidence items to act on first.
Bulk-select AI-recommended items
- After applying any filters you want, locate the Select [N] AI-Recommended button above the grid. The number in brackets reflects the count of flagged items in the current filtered view.
- Select the button. Every flagged item in the current view is selected and each row's Suggested qty field is pre-filled with the recommended order quantity.
- Review the pre-filled quantities. You can adjust individual quantities before proceeding to create purchase orders.
Important: Selecting Select [N] AI-Recommended does not place any orders. It only selects rows and pre-fills quantities so you can review them before creating purchase orders.
Act on individual rows
- Hover over any row in the action grid.
- Two controls appear on the row:
- Thumbs-up β confirms the recommendation looks correct. The system records this as a positive feedback signal.
- Thumbs-down β flags the recommendation as inaccurate. The system records this as a negative feedback signal for future model improvement.
- Create PO β opens the purchase order creation flow for that single product row.
- Select the appropriate control.
Note: If you want to provide more detail than a thumbs-down, open the explainability panel for the row and submit a quantity override with a reason code. This also feeds back into model training. See Review the explainability panel for a recommendation for details.
Give feedback on a recommendation
- Hover over the row you want to give feedback on.
- Select thumbs-up if the recommendation is accurate, or thumbs-down if it is not.
- The system displays Feedback submitted and records the signal.
Tip: Consistent feedback over time improves forecast accuracy for your center. Even a few weeks of thumbs-up/thumbs-down signals can meaningfully improve T3 and T4 products.
Worked example
Example β Filtering to urgent, high-confidence items:
An inventory coordinator at a medspa opens the action grid on a Monday morning. She applies Stock status: Critical and Confidence: T1 β High. The grid narrows to 4 products. She selects Select [4] AI-Recommended, which pre-fills quantities for all four rows. She reviews the suggested quantities β one product shows a suggested qty of 24 units at a vendor price of $8.50, giving a Value to order of $204.00. She adjusts the quantity to 20 to stay within budget, then proceeds to create purchase orders for the remaining three at the suggested quantities.
Tips
- Work Critical items first. The grid is pre-sorted so Critical products with the fewest days of stock cover appear at the top β you do not need to re-sort manually.
- Use the confidence tier to calibrate trust. For T3 β Low and T4 β Very Low products, open the explainability panel before accepting the suggested quantity. The panel shows the inputs and calculations so you can decide whether to override.
- Save time with bulk selection. Use Select [N] AI-Recommended after filtering to a manageable subset rather than selecting rows one by one. You can still adjust individual quantities after bulk-selecting.
- Use free-text search for spot checks. If a specific product is flagged by a provider, type the product name or SKU directly into the search field rather than navigating through category filters.
- Reset filters between sessions. If the grid appears to show fewer items than expected, check for active filter chips and select Reset all to confirm you are viewing the full prioritized list.
Troubleshooting
The action grid shows no rows.
- Confirm you are in center view, not organization or zone view. The action grid requires a specific center to be selected.
- Verify that the nightly refresh has run. If the module was activated today, the first recommendations will appear after the next nightly refresh.
- Check whether all filters are cleared. Select Reset all to remove any active filter chips that may be hiding rows.
The "Select [N] AI-Recommended" button shows 0 or is not visible.
- The button count reflects flagged items in the current filtered view. If filters are applied, some flagged items may be excluded. Select Reset all and check the count again.
- If no products have a Critical or Low status, there are no AI-recommended items to select. Check the Status column to confirm.
A row shows "--" for vendor price or value to order.
- This indicates no vendor is linked to the product. The grid displays -- when no vendor purchase-order history exists and no vendor is assigned on the product master. Navigate to the product's vendor settings in the Inventory module to assign a vendor, then wait for the next nightly refresh.
Feedback controls (thumbs-up / thumbs-down) do not appear when hovering.
- Ensure your pointer is directly over the row content, not on a blank area of the grid. The controls appear on hover within the row bounds.
- If the controls still do not appear, refresh the page and try again. If the issue persists, contact your administrator to confirm your access level.
Days of stock left appears unexpectedly high or low.
- Check the Daily usage value for the product. If daily usage is very low (near zero), days of stock cover will be very high and no reorder alert will fire β the product may be flagged as slow-moving instead.
- Check the Lead time column. If lead time was recently overridden, the new value takes effect at the next nightly refresh, not immediately.