Data Architect Agent

Setting up a Business View is normally a multi-step, multi-module workflow. You connect to a data source, profile the tables, configure joins in the data preparation interface, create calculated columns, enrich metadata, and validate the result. Each step has its own screen, and the process can take anywhere from 10 to 25 hours depending on schema complexity, particularly when derived tables, custom formulas, and thorough validation are involved.

Data Architect collapses this entire workflow into a single Kaiya conversation. It is a Kaiya agent that creates, edits, and validates Business Views conversationally from raw data sources. The agent handles schema analysis, join inference, calculated column creation, metadata enrichment, and pre-publish validation without requiring you to navigate between modules or write SQL.

Two modes of operation

Data Architect supports two starting points. You can create a new Business View from scratch by selecting tables from a connected data source, or you can open an existing Business View for editing. When editing an existing BV, the agent loads the current YAML configuration and makes it available for conversational modification, including adding tables, adjusting joins, creating columns, or running validation.

Snowflake and Redshift (in Live mode) are the only supported data sources for direct table discovery. For all other data sources, the datasets must be pre-loaded into Tellius before starting a Data Architect conversation.