Use of Taxonomy Augmented Generation to Improve Quality in Generative AI Solution

TAG to reduce Hallucination

This presentation was delivered at the DGIQ Data and AI Governance Conference in Washington, DC, in December 2024 by Ms. Neena Sathi (Applied AI Institute) and Dr. Arvind Sathi (KPMG).

It explores how Taxonomy Augmented Generation (TAG) enhances Generative AI by reducing hallucinations, improving domain-specific accuracy, and structuring AI-generated content. A case study on GradeBot demonstrates how TAG improves grading efficiency and accuracy. The results show a 20-30% accuracy boost and broader AI applications like automated grading, product recommendations, and structured decision-making. TAG is highlighted as a key approach for governing AI outputs and ensuring business reliability.

DGIQ Conference Presentation Final.pdf

Download PDF • 2.55MB