Today
Missing master data shows up only when it hurts: a shipment is stuck at customs because the tariff number is missing, a freight invoice is wrong because the weight field is empty, or a supplier declaration names the wrong country of origin. Nobody knows exactly who owns which field, and the big clean-up every few years lasts only a few months.
An example
At Nordhanse, 1,870 of 14,200 spare parts for reach stackers and spreaders lack a customs tariff number and 640 lack a weight. A spreader spare part for a port operator in Norway is stuck at customs for three days because of it. demi sends the gaps to the responsible product managers every morning, proposes 410 weights from drawings and shows after eight weeks that the share of complete export items has risen from 84 to 97 per cent.
Your data, wherever it lives
Your master data lives today in the ERP, the PLM, the CRM and product databases, plus spreadsheets that individual departments keep on the side, and in data sheets and supplier declarations as PDFs. demi reads, understands and links them per item, customer and supplier, with your rights. Every night the AI checks everything a clean-up project would take weeks for, and the owners decide which value is right.
Where the data lives today
- ERP
- PLM
- CRM
- Databases
- Office files
- PDFs
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
Data sources in this case: ERP, PLM, CRM and the product database, plus data sheets and supplier declarations. demi replaces nothing and links, per item, customer and supplier, the fields from every system with the documents a missing value can come from. It writes back only after approval.
- 2 · PermitPermission Layer
Each owner sees the error list of their area, management sees the rates. Prices and terms stay with purchasing and sales. Only people who maintain a field today may change it, and every change is recorded with person and source.
- 3 · ActIntelligence Layer
Every night the agent checks against rules you set: mandatory fields per item type, valid customs tariff numbers, plausible weights, identical values across systems. demi computes the rates exactly from the records. For missing values the agent proposes a value from a data sheet, drawing or supplier declaration, with its source.
How it runs
- 1
You define which fields are mandatory per item type, customer and supplier, and who owns them.
- 2
Every night demi checks all records in ERP, PLM and CRM against these rules and against each other.
- 3
Each area gets a rate, with its trend over the weeks and the costliest errors at the top.
- 4
Every owner gets their list in the morning, many gaps already with a proposed value and its source.
- 5
They confirm or correct, demi writes back and measures again the next morning.
What changes
Gaps show up before a shipment gets stuck
Clear ownership for every field
Data quality measured, not guessed
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
+180 hfreed per year
Assumptions and calculation
demi finds gaps and contradictions every night and proposes values with sources; you confirm.
15 min → 10.5 min per case
How we calculate
200 master data errors × 15 min × 30% × 1.0 (manual) × 12 months = 10,800 min = 180 h
Value per year, before demi costs
Freed per year+180 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
+€11,700
gained per year
180 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
What exactly does demi measure in the master data?
What you define: mandatory fields per item type, customer and supplier group, valid formats such as customs tariff number or VAT ID, plausible ranges such as weight per dimension, and identical values in every system. From the hits demi computes a rate per area, exactly from the records, not estimated.
Where do the proposals for missing values come from?
From your documents and systems: weights from drawings or weighing data, customs tariff numbers from similar items or earlier customs declarations, countries of origin from supplier declarations. Every proposal names its source. It is applied only when the owner confirms it.
Who is responsible for which errors?
You set that once, per field and area: weights go to product management, tariff numbers to the export team, payment terms to accounting. demi distributes the lists accordingly. If a list is left untouched, the head of that area sees it.
Does this replace a master data clean-up?
It turns the clean-up into a daily routine. A big clean-up tidies up once, and then the gaps grow back. demi checks every night, including every newly created item, so an error shows up the day after it was created, not at the first shipment.
