Google has added three new AI-powered SQL functions to BigQuery, aiming to make large-scale data analysis simpler for businesses. These functions—AI.IF, AI.CLASSIFY, and AI.SCORE—allow companies to use large language models (LLMs) directly within SQL queries, working on both structured data like numbers and unstructured data like text or images. The best part? They don’t require prompt tuning or extra tools, which used to make AI integration complicated and slow.

Why These New Functions Matter for Data Teams

In the past, using AI inside SQL was a real challenge. It involved moving data out of warehouses, doing complex prompt engineering, and manually choosing and tuning models. This process was time-consuming and costly. As Bradley Shimmin from The Futurum Group explains, SQL wasn’t designed to understand the nuance of unstructured data like customer reviews or support tickets. So, analysts often had to export data, wait for data scientists to process it, and then bring it back into SQL for analysis. That workflow could take days or even weeks.

Now, with these new AI functions, all that work can be done in a single SQL query. For example, AI.IF can filter or join data based on meaning, AI.CLASSIFY can categorize text or images, and AI.SCORE can rank data rows based on natural language criteria. This shift means analysts no longer need to learn complex prompt engineering or rely on external AI tools. Instead, they can just write standard SQL, making AI-driven insights faster and more accessible.

How These Functions Help Companies Save Time and Reduce Costs

One of the biggest advantages is that these functions are managed by Google. They handle all the technical heavy lifting behind the scenes—model selection, prompt optimization, and query tuning—so users don’t have to worry about it. Stephanie Walter from HyperFRAME Research points out that this makes AI more approachable for businesses. It lowers the skill barrier, meaning teams don’t need specialized AI experts to get started.

This managed approach also reduces operational risks. Instead of teams experimenting with different models and tuning parameters for each query, Google’s backend takes care of it. This not only speeds up the process but also cuts costs. Companies can get faster insights without the overhead of managing complex AI infrastructure.

The Growing Trend of AI in Data Warehousing

Google isn’t alone in adding AI functions to data warehouses. Other platforms like Databricks, Snowflake, and Oracle are also integrating AI capabilities. For example, Databricks offers AI functions for generative AI and LLM inference directly from SQL or Python. Snowflake has features for document parsing, semantic search, and AI-driven analytics. Oracle’s Autonomous Data Warehouse also supports AI workflows alongside SQL.

Phil Fersht from HFS Research sees this trend as part of a bigger shift. He believes these AI functions are the foundation for more autonomous data systems. Imagine AI agents inside data warehouses that can query, interpret, and make decisions in real time, all without leaving the platform. This evolution could transform how enterprises handle data—making their systems smarter and more self-sufficient.

Right now, these new functions are in public preview, so companies can try them out and see how they work. It’s clear that AI is increasingly becoming a core part of data analysis tools, shaping the future of enterprise data management.

Inspired by

• https://www.infoworld.com/article/4089869/google-bigquery-gets-managed-ai-functions-to-simplify-unstructured-data-analysis.html

Sources

• gmpg.org
• futurumgroup.com
• hfsresearch.com
• podigee.com
• podigee.com