Reference: Henkel Adhesive Technologies

Building a structured database through the automated extraction of technical data sheets from the global supplier network

Henkel AG & Co. KGaA, through its Adhesive Technologies business unit, is a leading global supplier of adhesives, sealants, and coatings. As part of a pilot project with Helm & Nagel GmbH, the company evaluated the automated extraction of technical data sheets from its international supplier network. To capture the relevant product properties and measurement values, the company used the Konfuzio software from Helm & Nagel GmbH.

Background

Technical data sheets from the international supplier network are received by Henkel Adhesive Technologies in a wide variety of formats and standards. Manually entering and harmonizing this diverse product information is time-consuming, error-prone, and ties up valuable resources within the department. Automated extraction can resolve this issue while simultaneously creating a structured, high-quality data foundation for downstream applications such as product consulting, supplier comparisons, or internal data management.

„The pilot project with Konfuzio helped us evaluate the technical approach for the automated extraction of our technical data sheets. Our collaboration with the team at Helm & Nagel GmbH was professional, constructive, and solution-oriented from the very beginning. We are very satisfied with how the project progressed and the results achieved.“

Stephan Künster, Senior Manager of Startup Innovation, Henkel Adhesive Technologies

Solution

Henkel, in collaboration with Helm & Nagel GmbH, evaluated the extraction pipeline of the Konfuzio software in a pilot project. The solution automatically extracts all relevant product information—such as product properties, measured values, and test data—from the technical data sheets and converts it into a structured, queryable data schema. Henkel’s subject-matter experts receive the extracted content in a structured format via the Konfuzio platform for further processing. The pilot project was carried out through close, iterative collaboration between the teams at both companies.

Pilot phase (Q1, 2026):

The first phase involved launching the project and adapting the extraction pipeline to the heterogeneous document base from Henkel’s international supplier network. To ensure reliable and high-quality data extraction, a flat data schema with explicit data types was implemented. In close collaboration with Henkel’s subject matter experts, the solution was developed through an iterative quality optimization process.

Extracted data:

• General Product Information: Product Name, Manufacturer, Document Metadata
• Material Properties and Measurement Values: Physical and Chemical Parameters with Values, Units, and Measurement Conditions
• Test Specifications: References to Relevant Test and Standard Requirements

Special features:

• Extraction from heterogeneous supplier formats without document-specific pre-training
• Schema-based validation with explicit value types for numbers, ranges, thresholds, and qualitative information
• Iterative quality improvement through automated feedback loops without adjusting model weights
• Human-in-the-loop interface for manual verification and correction, with a direct link between the value and its source in the document

Result

The pilot project achieved the jointly defined quality target across all tested supplier formats. The automated solution from Helm & Nagel GmbH provides a technically validated database for processing data sheets from heterogeneous supplier formats, thereby establishing a solid foundation for the project’s continuation.

Operational successes:

• Successful extraction from a selection of technical data sheets from the international supplier network
• Achieving the jointly defined quality goal across the entire pilot base
• Processing heterogeneous supplier formats without document-specific pre-training
• AI-powered extraction with schema-based validation

Strategic added value:

• Technical Validation: Verification of Extraction Quality Across Diverse Supplier Formats
• Methodological Framework: Specific Design Principles for Schema Design, Extraction Strategy, and Post-Processing
• Compliance: ISO 27001-certified platform and GDPR-compliant EU hosting
• Scalability: The pipeline can be extended to include additional document classes and languages

Profile: