Jan 22, 2026: AI_AGG and AI_SUMMARIZE_AGG (General availability)

AI_AGG and AI_SUMMARIZE_AGG, AI-powered aggregation functions that help you to analyze and summarize large volumes of text data efficiently at scale, are now generally available.

Unlike row-by-row functions, such as AI_COMPLETE, these functions are optimized for set-based aggregation, making them significantly more efficient for mass processing. Based on internal performance testing, AI_AGG delivers up to twice the throughput compared to AI_COMPLETE when aggregating large datasets.

  • AI_AGG reduces a column of text based on a natural language instruction, such as extracting common themes or issues across thousands of records.

  • AI_SUMMARIZE_AGG produces an overall summary of a text column as a whole, such as summarizing customer feedback at the product or business level.

Suggested use cases for these AI aggregation functions include:

  • Customer feedback and survey analysis: extract top themes, complaints, or praise.

  • Support ticket and incident review: summarize frequent issues and outcomes.

  • Topic discovery and modeling: automatically surface dominant topics or themes in text collections (e.g., product reviews, forum posts) using natural language prompts.

  • Executive summaries: generate high-level narratives from large unstructured text sources.

Both functions support GROUP BY, enabling summarization and aggregation by product, region, time period, or customer segment.

For more information, see Cortex AI Functions.

Language: English