# AI Inventory Control Tower — reviewer quickstart

Spend about five minutes exploring how a planner, buyer or inventory manager can move from an inventory issue to a specific next action.

This demonstration uses fictional materials, suppliers and transactions. Its planning date is **October 1, 2026**. The dates below belong to that example, rather than today's inventory. There are 100 materials across Chicago and Dallas; each material has separate stock and demand at each plant.

Prefer to watch first? Open the [narrated demo video](/walkthrough/). It includes a conversational AI-generated voiceover and on-screen explanations.

## Try these three situations

1. **Prevent a shortage.** Open [Inventory overview](/demo/), then [Precision ball bearing M12 — IP-1001, Chicago](/materials/?material=ITM-001-CHI). The recommendation asks the buyer to request **132 units from PO-4100-01 by October 10**, ahead of its October 23 arrival. The current purchase order contains enough units for the immediate gap, but its timing is too late. Review the stock projection and evidence.
2. **Adjust planning for supplier performance.** On that material, find **Supplier performance & stock levels**. Northline Industrial's eight completed deliveries took **22 days**, compared with an approved **10 days**. Review the proposed minimum, maximum and operating target. These are proposed policy changes requiring approval; they do not change an open purchase order's promised arrival.
3. **Review excess inventory.** Open [Pressure sensor M12 — IP-1005, Chicago](/materials/?material=ITM-005-CHI). There are **312 usable units** against a policy maximum of about **100.5 units**. The recommendation is to review **211.5 units** of excess after checking demand and commitments. The displayed dollar value is an estimated acquisition-cost exposure, not confirmed savings.

## Try following through

On the [Action board](/actions/), search a material number and select **Review**. Choose a status, enter a relevant note, and select **Save decision**. Changing the dropdown alone does not save. A comment has its own **Post comment** button. Acceptance records approval; completion records execution and requires an execution date. Neither proves a business outcome.

The material recommendation is available without an AI request. Open the private demo link from your host to enable **Generate AI recommendation** without a ChatGPT account. That button requests an additional review for the selected built-in sample material. Generation is optional and uses API credits; opening a material or reading a saved response does not generate another answer. Each link has a limited allowance and expiry. Keep your private link with the intended reviewer; sharing it shares its allowance. Access or usage limits may prevent a new request while the rest of the demo remains usable. Uploaded data is for local import practice; hosted AI review is limited to the built-in sample materials.

## Useful things to know

- **Sample data (Excel)** downloads the raw synthetic inputs used in the demonstration.
- Use only synthetic examples when trying spreadsheet imports. Do not upload company records or enter confidential information into this public review demo.
- Saved action changes belong to this browser session. They are not shared work assignments, and they may be lost when the session ends. Browser-local rejection notes may persist separately until removed. **Reset demo** restores the sample and clears its saved decisions, AI reviews and remembered rejection notes.
- The demo does not send emails, place orders, update SAP, or change a real material master.
- A material can serve production, accessories, transfers, kits, service, maintenance or direct fulfillment. Those uses share the same stock; they are not additional inventory.

When you finish, copy the [feedback questions](/reviewer-feedback-template.md) and return your answers to the person who shared the demo. No feedback is submitted automatically.