Agent 07 · Retail
InventoryIQ Agent
Keeps stock data clean and drafts purchase orders before items run out.
Agent PackDineDeskLangGraph
Found42
Fixed + verified42
Held for approval0
Recall1.00
One offline run on demo data with planted problems (recorded platform test, November 2026).
How InventoryIQ heals data, step by step
INData inKaggle anirudhchauhan/retail-store-inventory-forecasting-dat
01Detect42 problems found
02Diagnose4 problem types named
03Plan42 safe to fix · 0 need a person
04Fix + verify42 fixed and re-checked
05ReportPlain-English report and Control Tower record
- Client problem
- Stock-outs lose sales; bad inventory counts make reorder rules useless.
- What it does
- Fixes negative/impossible stock and missing sales, then predicts stock-out risk per SKU-store.
- How it heals
- Sets negative stock to zero with a recount task, imputes missing sales from the store-product median, drafts purchase orders for stock-out risks and flags overstock for markdown.
- Human gate
- POs above a value limit wait for the buyer.
- Typical buyer
- Multi-store retailers, restaurants, distributors
- Built with
- LangGraph self-healing loop
- Demo data
- Kaggle anirudhchauhan/retail-store-inventory-forecasting-dataset (or synthetic equivalent).
- Tables
- inventory ← Kaggle anirudhchauhan/retail-store-inventory-forecasting-dataset
- Target outcome
- Target: 20-30% fewer stock-outs on fast movers.
- Path to production
- DineDesk / Commerce Pack with POS export or API.
Issues detected on the demo data
| Issue | Risk | Proposed fix | Count |
|---|---|---|---|
| out_of_range_inventory_level inventory_level=-55 outside [0, None] | low | Set to 0 and create a physical recount task | 6 |
| missing_units_sold 'units_sold' is empty | low | Impute from the store-product median | 6 |
| duplicate_record Duplicate on record_id | low | Drop duplicate stock record | 5 |
| stockout_risk P0001 at S001: 2.0 days of cover | medium | Draft purchase order for 7 days of demand | 25 |
How we demo it
- Load retail inventory with planted errors.
- Run InventoryIQ: data repaired, PO drafts created.
- Open the PO drafts file.
Built on AWS
Runs in the platform container on EC2 Graviton (or on demand in Lambda), writes healed tables to the S3 lakehouse, uses Amazon Bedrock only to write explanations, and reports six metrics to CloudWatch.
Related agents
Step into DigiQuench.
Agents that heal your data.
Watch an agent find planted problems, fix the safe ones and wait for your approval on the rest. Thirty minutes, live.