Today
Inserts, cutters and mold components run low without anyone noticing. They are reordered when the bin is empty, and then a machine stands still for a part that costs next to nothing. Rush orders and express surcharges follow.
An example
At Norvelta, the planned orders for the next three weeks need 60 solid carbide ball-nose cutters Ø 6 mm for hard milling 1.2343 at HRC 52. The tool crib holds 25, enough until the middle of next week, and the lead time is 7 days. The agent proposes that purchasing orders today and shows the calculation behind it.
Your data, wherever it lives
Tool crib stock sits today in spreadsheets or in the dispensing system, consumption per order and machine in the MES, planned orders, open purchase orders and lead times in the ERP. demi reads, understands and links them per wear part and tool, with your rights. Every day the AI works out in seconds what the planned orders will use up and when a bin runs empty, instead of someone noticing at a stopped machine, and purchasing approves every order.
Where the data lives today
- Office files
- Databases
- MES
- ERP
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
The case needs data from the ERP, MES and spreadsheets. demi replaces nothing and links tool crib stock, consumption per order and machine, planned orders and supplier lead times; every wear part knows its tool life, consumption and source. demi writes back only after approval.
- 2 · PermitPermission Layer
Tool crib staff and buyers see the proposals. Purchasing approves orders, and every approval is recorded.
- 3 · ActIntelligence Layer
The agent computes demand strictly from planned orders and recorded consumption rates and sets stock and lead time against it. If a gap looms, it proposes the order with the calculation behind it.
How it runs
- 1
demi reads the planned orders for the coming weeks and the consumption per operation.
- 2
The agent computes demand per wear part and tool.
- 3
Stock, open purchase orders and lead time are set against it.
- 4
If a gap looms, a reorder proposal with its reasoning goes to purchasing.
- 5
Purchasing approves, and only then does the order go into the ERP.
What changes
Fewer stops for missing wear parts
Fewer rush orders
Leaner tool stock
What does it bring you?
Check it with your own numbers
Every starting value is an assumption you can change. The calculation sits right below, step by step.
Your numbers
Result
+80 hfreed per year
Assumptions and calculation
demi tracks usage per machine and order and proposes the reorder; you place it.
10 min → 6 min per case
How we calculate
100 reorders × 10 min × 40% × 1.0 (manual) × 12 months = 4,800 min = 80 h
Value per year, before demi costs
Freed per year+80 h
Assumption: labor cost of $48.62 per hour in US manufacturing in June 2026 (BLS), plus workplace, IT and overhead.
= Value per year, before demi costs
+$5,200
gained per year
An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.
Combine with more cases and get the packageQuestions about this use case
Where does the consumption data come from?
From withdrawals at the tool crib, confirmations in the MES or existing spreadsheets. Where tool life is recorded per tool, the agent uses it directly. Where it is not, it works from withdrawals per order.
Does demi take planned orders into account?
Yes, that is the core of it. Demand comes not only from past consumption but from what will be produced in the coming weeks. When the plan changes, the proposal changes.
Who approves the orders?
Purchasing. demi orders nothing itself; it puts forward a proposal with quantity, supplier and reasoning. demi writes back only after approval.
What if the stock in the system is wrong?
That usually shows when withdrawals and stock are compared. The agent flags parts with implausible stock and asks the tool crib for a count before it proposes an order.
