Hyland Knowledge Enrichment

Unstructured data makes up as much as 80% of enterprise content and is scattered across an average of 21 repositories. Unstructured data includes documents, emails, images, audio files and more. Unlike structured data in databases, unstructured content is inconsistent and harder to search, analyze or use in business processes and AI applications.

Knowledge Enrichment uses data curation techniques to structure and normalize unstructured content, making it consistent and ready for use. Powered by Document Filters it extracts and transforms data from over 600 file formats while preserving the original context and logical structure. This allows content to remain meaningful and usable across downstream AI, analytics, and automation applications.

Knowledge Enrichment processes over 600 file formats, including common business documents (PDF, Word, Excel), multimedia files (images, audio, video), emails, scanned documents, code, markup and specialized formats like CAD files (DWG, DGN, STEP, DWF). Most of these formats are provided through Document Filters, with a list located in the documentation.

Unlike many competitors, Knowledge Enrichment uses a combination of both LLMs and deterministic techniques like Document Filters for precision extraction. This approach delivers superior accuracy because it preserves document structure and positional context while reducing hallucination risks. Tables stay tables, headers stay headers, and you get ready-to-use outputs without additional cleaning or formatting. This deterministic extraction method allows consistent, reliable results that you can trust for enterprise applications.

Knowledge Enrichment includes built-in PII masking capabilities that identify and protect sensitive information (names, emails, addresses, social security numbers, account numbers) across all supported file types. Configurable policies let you decide what to mask or preserve based on your compliance requirements.

No. Knowledge Enrichment goes far beyond traditional metadata tagging. It performs contextual text chunking, preserves positional information, generates semantic embeddings, identifies relationships between documents and extracts meaningful entities while maintaining document structure. The result is content that's not just tagged but truly understood by AI systems. This enables advanced capabilities like semantic search, intelligent recommendations and context-aware automation that simple metadata can't support.

Yes, Knowledge Enrichment is designed for builders — app developers, data engineers and solution builders. It’s an API-driven solution designed to integrate into broader architectures and workflows but with options catering to varying needs.

• The data curation and context enrichment capabilities through the Knowledge Enrichment API offers faster time-to-value for organizations that want enriched, contextual output without dealing with the complexity of extraction.
• Document Filters, a proven Hyland technology powering Knowledge Enrichment’s data curation, is perfect for technical teams and organizations looking for full control over how content is enriched, structured and delivered for downstream use.

Hyland provides AI-readiness offerings that cater to your organization’s needs and technical resources.

• For organizations that want enriched, contextual output without managing complex extraction, the Knowledge Enrichment API comes with both data curation and context enrichment capabilities. This is the perfect solution for organizations that have use cases around AI-ready data, domain-specific linking, or connecting concepts across documents.
• For organizations and solution builders that want full control while extracting structured text and metadata from a wide range of file types, Document Filters is the right option. This gives you the ability to perform data curation in your own infrastructure so you can build custom workflows that require normalized content.

Ease of implementation will depend on the needs of your organization.

The Knowledge Enrichment APIs enable organizations to focus on preparing content for knowledge graphs, intelligent retrieval, RAG pipelines or industry-specific LLM workflows without dealing with the complexity of extraction.

Document Filters enables organizations to have full control over how extraction is done and how the data is integrated into their own infrastructure, including on-premises or offline systems. This process may entail additional work.

Knowledge Enrichment is currently available as a cloud-native API designed for seamless integration into modern data pipelines and enterprise architectures. For organizations requiring on-premises deployment, Document Filters offers full control over extraction and enables deployment in your own infrastructure, including on-premises or offline systems.

Yes. Knowledge Enrichment delivers AI-ready output in standardized formats (like JSON) that integrate seamlessly with any AI system, analytics solution, data lake or third-party tool. The API-first design ensures you can use enriched content wherever you need it — whether that's feeding external LLMs, populating data catalogs, powering RAG pipelines or integrating with custom applications.

Knowledge Enrichment and Knowledge Discovery work together to transform how you access and use enterprise content. Knowledge Enrichment prepares your content by converting unstructured data into structured, contextualized, AI-ready data. Knowledge Discovery then leverages that AI-ready data to power AI-driven search and natural language question answering, delivering faster, more accurate responses. When used together, Knowledge Enrichment improves the quality of search results and AI-generated answers in Knowledge Discovery.