Know which part will run short, and when.
The predictor reads every open order, its BOM, current stock, open purchase orders and real supplier lead times. It tells purchase which part will run short, on which date, which order it delays, and what to order today.
- 1
Shortage date per part
Projected stock is calculated day by day for every component. The first day it goes below zero is the shortage date.
- 2
Order impact
Each alert names the sales orders that will slip and by how many days, so the most urgent shortage is handled first.
- 3
Suggested purchase
A quantity and a supplier are suggested from lead time history, rate and delivery record. Purchase can change either.
F
Purchase / AI Shortage Predictor / Alerts
Forecast - BO-015 Mechanical seal
Runs short on 24 Oct. Delays SO-3318 by 6 days. Suggest 220 from Supplier B today (4 day lead time).
Based on
- 4 open sales orders - 610 pumps
- BOM FG-1001 Rev C - 1 seal per pump
- Stock 40 in Main store, 0 in transit
- Supplier B lead time 4 days (last 9 POs)
- 4
Reasons you can check
Every alert lists the orders, BOM revision, stock and lead times it used. Planners can open each one.
- 5
Real lead times
Lead time is learnt from actual PO to GRN dates per supplier and item, not from the number on the supplier form.
- 6
People approve, always
The predictor never places an order. A suggestion becomes a PO only when someone with purchase rights approves it.
Section A-A
How a prediction is made
- 01
Read demand
Open sales orders and production plan.
- 02
Explode BOMs
Released revision, scrap allowance included.
- 03
Net supply
Stock, open POs, WIP and vendor stock.
- 04
Apply lead times
Learnt from each supplier history.
- 05
Alert and suggest
Date, impact, reason and suggested PO.
Specification
What is in AI Shortage Predictor
| Inputs | Sales orders, production plan, released BOMs, stock by store, open POs, WIP, vendor stock |
|---|---|
| Lead time model | Actual PO to GRN days per supplier and item, with delay trend |
| Horizon | Day by day projection for the next 60 days |
| Refresh | Recalculated whenever an order, stock movement or PO changes |
| Output | Shortage date, quantity short, affected orders, delay in days, suggested PO |
| Control | Suggestions only. Purchase orders need human approval |
Notes
Common questions
- 01
Does the AI place orders?
No. It suggests a quantity, supplier and date. A person in purchase or planning reviews it and approves, edits or rejects it.
- 02
How do we know a prediction is right?
Each alert shows the data behind it: the orders, BOM, stock and lead times used. If any input is wrong, fixing it updates the prediction.
- 03
How much history does it need?
It works from day one with the lead times you enter. As POs and GRNs are recorded, it switches to actual supplier lead times.
Engineering change note - ECN-2026 - Your plant
Approved for production.
Change requested: move the plant from spreadsheets and separate books to one connected ERP. Here is the revision plan a single-plant rollout usually follows.
| Rev | Description | When |
|---|---|---|
| A | Masters, opening stock and ledgers imported from Tally and Excel | Week 1-2 |
| B | Sales orders, GST invoices and purchase running live | Week 2-4 |
| C | BOMs released, work orders and job cards on the floor | Week 4-6 |
| D | Shortage predictor switched on with real lead times | Week 6-8 |
Requested by
Plant owner
Reviewed by
Orrsap team
Approved
For production
Book a demo with our team.
We run it on your products, your BOMs and your parts list.