Automating Taxonomy Generation With Compound AI on Databricks

Taxonomy generation is a challenge across industries such as retail, manufacturing and e-commerce. Incomplete or inconsistent taxonomies can lead to fragmented data insights, missed monetization opportunities and stalled revenue growth. In this session, we will explore a modern approach to solving this problem by leveraging Databricks platform to build a scalable compound AI architecture for automated taxonomy generation.

The first half of the session will walk you through the business significance and implications of taxonomy, followed by a technical deep dive in building an architecture for taxonomy implementation on the Databricks platform using a compound AI architecture.

We will walk attendees through the anatomy of taxonomy generation, showcasing an innovative solution that combines multimodal and text-based LLMs, internal data sources and external API calls. This ensemble approach ensures more accurate, comprehensive and adaptable taxonomies that align with business needs.