Technical content may describe dimensions and material in prose.
Product systemP3 · transformation fieldOperationraw → structuredOutputattribute · value · unitEvidenceAC820825
Attribute extraction
Turn Unstructured Product Information into Structured Attributes
Technical information often exists — just not in a format your catalog can use.
A product page may mention dimensions inside a paragraph. A supplier table may contain specifications under inconsistent labels. Documentation may describe materials and properties without mapping them into your PIM schema.
ENRIVAQ converts this information into structured product attributes.
AC820825REFERENCE PRODUCT · SAME-SKU EVIDENCE
As foundUNSTRUCTURED SOURCE
Outer diameter: AC820825; inner diameter: 8716106986118; width: Optima / Optima HD
ENRIVAQEXTRACT
Canonical recordSTRUCTURED OUTPUT
AttributeValueUnitStatus
{{ a.attr }}
{{ a.val }}{{ a.unit }}structured01
Input
Useful product data can be hidden almost anywhere
Attribute extraction can begin with unstructured or semi-structured information such as:
Source terminology may use inconsistent labels for the same attribute.
Outer diameter: AC820825; inner diameter: 8716106986118; width: Optima / Optima HD
For example
OEM AC820825; EAN 8716106986118; application Optima / Optima HD
contains useful technical data, but it is still just text until the values are separated into structured fields.02
Extraction
Convert text into attribute, value and unit
The core output follows a simple structure: Attribute → Value → Unit
RAW TECHNICAL SOURCE
“Outer diameter AC820825; inner diameter 8716106986118; width Optima / Optima HD”
EXTRACTION FIELD
AttributeValueUnit
{{ e.attr }}
{{ e.val }}{{ e.unit }}This format allows downstream systems to work with the information as product data rather than as a sentence.
The exact attributes depend on the product category and target schema.
03
Mapping
Fit extracted data into the catalog you already use
Extraction is only useful when the output maps to the structure required by the business.
Source terminology{{ m.label }}
MAP
Canonical attributeOuter diameterOne field in the approved schema
The enrichment workflow needs to map source terminology into the approved attribute structure instead of creating a new field for every spelling variation.
04
Units
Keep values comparable across products
Different sources may use different units or formats. For example:
{{ u.src }}{{ u.val }}
SCHEMA-CONTROLLED OUTPUT42 mm
These may represent the same physical value, but they cannot be treated as two unrelated specifications.
Normalization allows the catalog to store technical values consistently. The target unit should be determined by the catalog schema.
05
Technical example
From technical text to a usable spare-part record
Use a real spare part here.
SOURCE INFORMATIONPublic source evidence: OEM reference · EAN · product type · compatibility · material
Extracted attributes
{{ r.k }}{{ r.v }}
Visual: source on the left, structured record on the right. This is stronger than describing extraction conceptually.
Validation
Extraction should not automatically mean acceptance
A value can be extracted correctly from a source and still be unsuitable for the product record. Possible issues include:
These cases should move into validation or review rather than being silently published.
Extracted but not accepted{{ valLabel }}
{{ v.name }}{{ v.tag }}review
Turn one unstructured product into a structured record
Send a representative product or technical source and see what can be extracted.