Overview
The Classifier node assigns meaning, structure, and business context to every document flowing through your system. It works across pages, full documents, and multi-document cases, leveraging AI Coach–trained models or the Reader + Agent for more advanced decision-making. It operates in two modes: Decision Maker (makes business-specific judgments based on content, quality, annotations, or rules) and Taxonomy Assigner (applies one or multiple labels according to your hierarchical taxonomy). The Classifier connects directly to routing, enrichment, validation, and review workflows, ensuring every document is accurately interpreted, verified, and sent to the correct destination.
Capabilities
-
AI Coach Integration
Train specialized, compact models tuned to your domain using example-based learning.
-
Dynamic Taxonomy Tracking
Automatically update labels, detect new emerging categories, and notify teams to evolve your taxonomy.
-
Confidence Thresholds
Trigger human review in the Reviewer node for items below calibrated confidence levels.
-
Landmark & Barcode Detection
Classify using tags, barcodes, or visual markers generated by the Tagger or embedded in documents.
-
Routing & Validation
Automatically send documents or cases to the right departments, people, or workflows based on their labels, while ensuring all required documents exist in assembled cases.
Benefits
-
High-Accuracy Classification
AI models trained on your data and rules deliver reliable results at scale.
-
Better Decision Workflows
Drive automated routing, approvals, and enrichment based on accurate labels.
-
Human Oversight When Needed
Confidence thresholds ensure experts review the edge cases.
-
Self-Evolving Taxonomy
Automatically surfaces new categories, keeping your business classification up to date.
-
End-to-End Traceability
Every classification decision is logged, explainable, and auditable.
FAQ
Yes. It supports multi-label classification aligned with hierarchical and overlapping taxonomies.
Changes are automatically tracked. The Classifier updates labels and alerts you if new categories should be added.
No. AI Coach allows you to train efficient models using small curated sets of examples.
Through confidence thresholds, fallback logic, and triggering human review for ambiguous items.
Yes. It can detect landmarks, barcodes, and Tagger-generated markers.
Absolutely. Labels can drive routing rules to departments, workflows, decision engines, or external systems.