This space contains a two-part project template for BAPA (Brand AI Perception Analytics). It is designed to demonstrate how to build an automated, cost-controlled multi-LLM data pipeline using KNIME and Google BigQuery.

Instead of processing rows locally, the workflows use in-database processing to handle delta loading, ensuring you only ever send new data to the LLM APIs. The generated brand awareness insights may be joined with existing digital analytics stored on BigQuery (i.e. Google tags) for further analysis without unnecessary data replication.

Workflows included:

• BAPA01 Ingest Gemini to Bronze

• BAPA02 Bronze to Silver Enrichment

BAPA01 is designed as a template built with Google's Gemini LLM service, which can be applied to other LLM services such as Claude, ChatGPT, Perlexity or Copilot.

For a full strategic breakdown on why this visual blueprint was built, database schemas and how to use it for Generative Engine Optimization (GEO), check out this detailed post here.