Others classify. We cite the source for every value.
From unstructured catalog to ETIM-ready data — automated, source-cited, in 24 hours. As part of the Zander Group we are an official ETIM partner with direct access to the ETIM API — no tool outage from API cutoffs. BMECat export, Excel list or PDF datasheet in. ETIM class with confidence, populated features with EV codes, ~80-character snippet for every numeric value out.
Class match (1k blind sample)
Fill rate vs. manufacturer
Chars of source proof per value
1,000 products classified
Built on proven standards and leading AI infrastructure
Standards
ETIM 9 / 10 · BMECat V5.0 · ETIM xChange V2.0 · GS1
AI infrastructure
Anthropic Claude · OpenAI · Google Gemini · Perplexity Sonar · OpenRouter
Certifications
Official ETIM partner · Zander Group
Confidence routing
What happens when the tool is uncertain?
≥ 90%
Auto-confirmed, ready for export
Spot-check review is enough
60–89%
Flagged as “Review”, with alternative suggestions
Manual review recommended
< 60%
“Insufficient data” — additional web research + datasheet enrichment required
Add data sources or decide manually
You stay in full control. The tool suggests, it does not decide. The audit trail shows the rationale, all alternatives and the sources used for every classification.
Standards & exchange formats
Compatible with every ETIM and exchange format — import and export.
Live example
Real Mitegro catalog, classified with source proof
10 products from 9 different ETIM classes, each with ETIM class, confidence, 15 populated features and a ~80-character source snippet per numeric value. Mitegro has approved publication.
How it works
From raw file to source-backed ETIM classification — four stages.
Data sources
BMECat V5, Excel/CSV, PDF datasheets, manufacturer pages, REST API. Multiple sources per product get combined.
Classification
A multi-stage workflow proposes the ETIM class and validates every suggestion against the official ETIM database. Fuzzy matching on invalid codes. Confidence score 0-100 per product.
Filling with proof
Per EV code: value + ~80-char snippet from datasheet, BMECat or web research. Hallucination filter discards unsourced values.
Validation + export
Validator filters values outside ETIM allowed lists. Backfill recovers rejected values without new AI calls. Export as BMECat V5, ETIM xChange V2 or Excel.
The pipeline
ETIM versions
ETIM 8, 9, 10 — plus the 9-to-10 migration competitors don't have ready.
Voices from the field
Three perspectives — already in production at the Zander Group.
We ran 1,000 products from our catalog through the classifier. For 93 percent, the ETIM class assignment matched our manual one. For the remaining 7 percent, the AI choice was in several cases more precise than our historical categorisation — we corrected our own master data based on the run.
Head of Product Data Management
Lighting manufacturer, Top 3 DACH
Until now we had no ETIM data — as a small company the effort to build it in-house was simply too high. With the tool we can finally deliver ETIM-compliant datasheets to our B2B wholesale customers without hiring a dedicated role.
Managing Director
Specialist electrical-installation manufacturer, < 50 employees
We use the tool internally at the Zander Group for classification and enrichment of additional ETIM features — the output matches the quality of our specialists but is multiple times faster. That is exactly what we built it for.
Head of Product Data
Zander Group (internal use)
Case study
Pilot at Mitegro Germany — 500 products in one run, fully source-backed
Starting point
Mitegro Germany aggregates electrical products from many manufacturers as a wholesaler — from lighting via switchgear to HVAC. Data structures, datasheet formats and detail depth differ heavily between manufacturers. Manual ETIM classification of a mixed catalog ties up significant headcount and does not scale linearly with portfolio growth.
Setup
A first Mitegro sample with 500 products across 35 manufacturers (CMD, HT-Instruments, INDEXA, ABB, sonnen, Finder, Siemens, Dietzel, Cosmo, Daikin and others) was handed to the ZDS pipeline. The tool autonomously researched manufacturer datasheets per product, parsed PDFs and webpages and enriched master data from the BMECat feed.
Outcome
All 500 products were classified, average confidence 88 percent, on average 15 source-backed features per product — about 7,500 values in total. Every numeric value carries a ~80-character source snippet from the original. Class match against Mitegro master data was 88 %; the 12 % deviations were flagged in the audit trail and handed back for visual review.
Mitegro statement
“Manual ETIM classification would have been a multi-day effort. With the ZDS tool we were through in under 8 hours — with better source depth than our existing data feeds and an audit trail we can push back to our suppliers.”
Example: INDEXA surveillance camera with manufacturer datasheet as data basis
See the full Mitegro demo →Frequently asked
What procurement, data and compliance ask first.
Pricing
What does it cost?
We work project-based with no monthly subscription. The per-product price drops with order size:
100 products (minimum order)
from €0.80
1,000 products
from €0.40
10,000 products
from €0.20
✓ Plus one-time setup fee for style calibration, data integration and delivery-format definition (€500–3,000 depending on complexity).
★ Pilot program: 100 products classified free of charge within 24 hours — you see classification, fill rate and audit trail before deciding.
For a concrete quote on your catalog: a brief requirements call, then a tailored estimate.
100 products in 24 hours.
Free.
Send us a small slice of your catalog. We return classification, populated features and audit trail — fully reviewable. Only after your green light do we discuss a paid engagement.
Request the pilot