Skip to main content

529-Type Ontology

Talonic classifies every document using a 529-type ontology — a hierarchical taxonomy of document categories that guides field discovery, schema suggestion, and extraction accuracy.

Ontology structure

The ontology organizes 529 document types into a three-level hierarchy: domain, category, and type. During the Resolve phase of the four-phase pipeline, the system classifies each document and uses the classification to select appropriate Field Registry entries.

LevelCountExamples
Domain12Finance, Legal, Logistics, Healthcare, Energy
Category67Accounts Payable, Contracts, Shipping, Clinical
Type529Invoice, Purchase Order, Bill of Lading, Lab Report

DIN SPEC 91491 alignment

The ontology aligns with DIN SPEC 91491, the German standard for AI-assisted document processing that Talonic co-authored. Domain and category levels map directly to the standard's document classification scheme. This ensures interoperability with other DIN SPEC 91491-compliant systems.

Auto-classification

When you submit a document to the extract endpoint without specifying a schema, the pipeline uses the ontology to classify the document and suggest relevant fields. The classification confidence is included in the extraction data response.

Schema Graph integration

The Schema Graph links schemas to ontology types. A schema associated with the "Invoice" type inherits default fields from the ontology, such as vendor_name, invoice_number, and total_amount. Use the search API to query documents by ontology type.

Custom types

Enterprise customers can extend the ontology with custom types specific to their domain. Custom types integrate with the Field Registry tier system — custom type fields are assigned Tier 3 (Custom) status. Contact Talonic for Enterprise ontology customization.