New tool automates ontology design, entity extraction and knowledge graph generation

VAIV Company has unveiled an ontology automation solution designed to automatically build ontology-based knowledge graphs from enterprise documents and deliver evidence-backed AI responses. The solution will be integrated into the company's VAIV Ontology RAG Platform.

Ontology technology structures data around concepts and their relationships, helping reduce hallucinations in generative AI. Most enterprise knowledge assets, including product terms and conditions and user manuals, exist as unstructured documents. Converting those documents into ontologies has traditionally relied on domain experts, making the process both time-consuming and costly while offering limited visibility into the underlying design process.

VAIV Company said the new solution makes the ontology development process transparent and provides an intuitive workflow that enables business users with domain expertise to participate directly. Once documents are uploaded, AI automatically proposes an initial ontology schema, which users can immediately revise and finalize through a conversational text interface. The company said the approach significantly reduces reliance on specialists during the initial data modeling stage.

Once the schema is finalized, the solution automatically extracts key information (entities) and relationships from documents and instantly visualizes them as a knowledge graph. According to VAIV Company, the system also shows the reasoning path behind AI-generated answers by identifying which concepts within the knowledge graph were used as supporting evidence.

For example, when an insurance policy document is uploaded and a user asks about exclusions or benefit reduction conditions, the platform connects related clauses scattered throughout the document into a unified relationship graph and provides accurate answers together with references to the original source.

View other news Copyright © The Elec Inc. No unauthorized transfer and redistribution