dbt Summit
All on demand sessions
On demand session

SQL-first AI: bringing BigQuery AI into your dbt project

Google Cloud on wiring BigQuery's AI SQL functions into dbt pipelines, from sentiment scoring to LLM-generated product briefings.

Recorded at dbt Summit · 2026Breakout session archive
01

Overview

Breakout session · 2026

Demonstrates integrating BigQuery AI functions directly into dbt projects (AI.GENERATE, AI.CLASSIFY, AI.SCORE) to turn unstructured logs into actionable insights within SQL workflows, covering cost optimization, data quality via dbt tests, and managing LLM latency.

What you'll take away

  1. 01

    BigQuery's AI.GENERATE and its managed functions (AI.IF, AI.CLASSIFY, AI.SCORE, AI.AGG) let a dbt model call an LLM directly from a SQL SELECT statement, no separate Python pipeline required.

  2. 02

    The first rule for production: never run an AI function inside a full refresh — wrap the logic in an is_incremental() block so only new rows get sent to the model.

  3. 03

    Pairing BigQuery's output_schema parameter with standard dbt tests like not_null and accepted_values catches bad AI output in staging before it reaches a mart.

Return to all sessions
02

Session speakers

02 speakers

  • Alicia Williams

    Google

  • Jobin George

    Google