Today
System stock and physical stock drift apart and nobody knows why. When material is missing, people go looking, production waits and so does the customer. The stocktake fixes the number but not the cause.
An example
At Talwerk Antriebstechnik the system shows 400 M8 × 25 hex bolts, the bin holds 150. The agent finds two production orders for geared motors that used the bolts without the backflush being booked. It proposes the correcting entry and the warehouse lead approves it.
Your data, wherever it lives
Stock and bookings sit in the ERP today, bins and movements in the WMS, production confirmations in the MES and inventory counts in spreadsheets or on paper lists. demi reads, understands and links them per article and bin, with your rights. The AI explains a discrepancy in minutes instead of a long search in the warehouse, and the warehouse lead decides on the correction.
Where the data lives today
- ERP
- MES
- Office files
- Scans & paper
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, WMS, MES and spreadsheets. demi replaces nothing and brings together warehouse movements, goods receipts, issues, production confirmations and counts per article and bin. demi writes back only after approval.
- 2 · PermitPermission Layer
Warehouse leads see their own areas. The warehouse lead approves each correction, and it is recorded with cause and evidence.
- 3 · ActIntelligence Layer
The agent calculates discrepancies exactly from book stock, count and movements and explains typical causes such as a missing backflush or a wrong unit of measure. Every explanation cites the bookings it rests on and ends in a proposed correction.
How it runs
- 1
demi reconciles system stock with counts and movements.
- 2
The agent flags articles and bins with a discrepancy.
- 3
It traces the bookings back and names the likely cause.
- 4
The warehouse lead reviews the proposal and approves the correction.
- 5
The correction is booked and the cause goes to whoever can fix it.
What changes
Fewer stock surprises
Causes fixed, not just numbers corrected
Availability you can rely on
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
+160 hfreed per year
Assumptions and calculation
demi compares postings, movements and counts and explains the likely cause; you correct it.
20 min → 12 min per case
How we calculate
100 stock discrepancies × 20 min × 40% × 1.0 (manual) × 12 months = 9,600 min = 160 h
Value per year, before demi costs
Freed per year+160 h
Assumption: labour cost of €49.50 per hour in German manufacturing in 2025 (Destatis), plus workplace, IT and overheads.
= Value per year, before demi costs
+€10,400
gained per year
160 h ÷ 1,760 h per role = 0.1 full-time roles
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
How does demi find the cause of a discrepancy?
The agent goes through every movement of the article since the last count that matched. It looks for known patterns: consumption without confirmation, transfers to the wrong bin, pieces booked as cartons, duplicate goods receipts. For each suspicion it shows the bookings so you can check them yourself.
Does demi book corrections on its own?
Not without approval. The agent prepares the correcting entry, the warehouse lead confirms it, and only then does the result go back. demi writes back only after approval. Cause, approval and booking sit together in the record.
Does it work with our warehouse system?
Your data sits in the ERP, WMS and MES. The data layer is built for them; how the data gets to demi, we clarify in a conversation. If you have no separate WMS, the stock bookings in the ERP and count sheets in spreadsheets are enough. Nothing is replaced.
Do we need a stocktake for this?
No. The agent also checks continuously for inconsistencies such as negative stock or issues without a receipt. Every count, whether a spot check, cycle count or year-end stocktake, sharpens the picture.
