SQL Keywords

AI_SUMMARIZE_AGG

Feb 23, 2026

min read

AI_SUMMARIZE_AGG

Overview

Aggregates a text column and returns a summary across multiple rows. Not subject to context window limitations.

Syntax

AI_SUMMARIZE_AGG(text_column)

AI_SUMMARIZE_AGG(text_column)

AI_SUMMARIZE_AGG(text_column)

Parameters

text_column (VARCHAR): Column containing text to summarize

Use Cases

Summarize customer feedback by product
Create executive summaries from detailed records
Generate meeting highlights across multiple sessions
Consolidate survey responses
Aggregate news articles by topic

Code Examples

Example 1: Summarize Reviews by Product

SELECT 
    product_id,
    COUNT(*) AS review_count,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM customer_reviews
GROUP BY

SELECT 
    product_id,
    COUNT(*) AS review_count,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM customer_reviews
GROUP BY

SELECT 
    product_id,
    COUNT(*) AS review_count,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM customer_reviews
GROUP BY

Output:

product_id | review_count | summary
-----------|--------------|--------
101        | 45           | Customers appreciate the battery life and build quality. Main concerns are price and weight.
102        | 32           | Highly rated for performance. Users report occasional connectivity issues

product_id | review_count | summary
-----------|--------------|--------
101        | 45           | Customers appreciate the battery life and build quality. Main concerns are price and weight.
102        | 32           | Highly rated for performance. Users report occasional connectivity issues

product_id | review_count | summary
-----------|--------------|--------
101        | 45           | Customers appreciate the battery life and build quality. Main concerns are price and weight.
102        | 32           | Highly rated for performance. Users report occasional connectivity issues

Example 2: Monthly Meeting Summaries

SELECT 
    DATE_TRUNC('month', meeting_date) AS month,
    AI_SUMMARIZE_AGG(meeting_notes) AS monthly_summary
FROM team_meetings
WHERE meeting_date >= CURRENT_DATE - 180
GROUP BY month
ORDER BY month

SELECT 
    DATE_TRUNC('month', meeting_date) AS month,
    AI_SUMMARIZE_AGG(meeting_notes) AS monthly_summary
FROM team_meetings
WHERE meeting_date >= CURRENT_DATE - 180
GROUP BY month
ORDER BY month

SELECT 
    DATE_TRUNC('month', meeting_date) AS month,
    AI_SUMMARIZE_AGG(meeting_notes) AS monthly_summary
FROM team_meetings
WHERE meeting_date >= CURRENT_DATE - 180
GROUP BY month
ORDER BY month

Example 3: Customer Feedback by Region

SELECT 
    customer_region,
    AI_SUMMARIZE_AGG(feedback_text) AS regional_feedback
FROM customer_feedback
WHERE feedback_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    customer_region,
    AI_SUMMARIZE_AGG(feedback_text) AS regional_feedback
FROM customer_feedback
WHERE feedback_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    customer_region,
    AI_SUMMARIZE_AGG(feedback_text) AS regional_feedback
FROM customer_feedback
WHERE feedback_date >= CURRENT_DATE - 30
GROUP BY

Example 4: Summarize Support Tickets

SELECT 
    ticket_category,
    COUNT(*) AS ticket_count,
    AI_SUMMARIZE_AGG(ticket_description) AS category_summary
FROM support_tickets
WHERE status = 'Resolved'
  AND resolved_date >= CURRENT_DATE - 7
GROUP BY

SELECT 
    ticket_category,
    COUNT(*) AS ticket_count,
    AI_SUMMARIZE_AGG(ticket_description) AS category_summary
FROM support_tickets
WHERE status = 'Resolved'
  AND resolved_date >= CURRENT_DATE - 7
GROUP BY

SELECT 
    ticket_category,
    COUNT(*) AS ticket_count,
    AI_SUMMARIZE_AGG(ticket_description) AS category_summary
FROM support_tickets
WHERE status = 'Resolved'
  AND resolved_date >= CURRENT_DATE - 7
GROUP BY

Example 5: Sales Call Summaries

SELECT 
    sales_rep_id,
    sales_rep_name,
    AI_SUMMARIZE_AGG(call_notes) AS weekly_summary
FROM sales_calls
WHERE call_date >= DATE_TRUNC('week', CURRENT_DATE)
GROUP BY sales_rep_id,

SELECT 
    sales_rep_id,
    sales_rep_name,
    AI_SUMMARIZE_AGG(call_notes) AS weekly_summary
FROM sales_calls
WHERE call_date >= DATE_TRUNC('week', CURRENT_DATE)
GROUP BY sales_rep_id,

SELECT 
    sales_rep_id,
    sales_rep_name,
    AI_SUMMARIZE_AGG(call_notes) AS weekly_summary
FROM sales_calls
WHERE call_date >= DATE_TRUNC('week', CURRENT_DATE)
GROUP BY sales_rep_id,

Data Output Examples

Product Review Summary

Input: 50 customer reviews
Output: "The product receives high marks for design and functionality. 
Customers frequently mention the intuitive interface and fast performance. 
Common complaints include limited color options and higher price compared to competitors

Input: 50 customer reviews
Output: "The product receives high marks for design and functionality. 
Customers frequently mention the intuitive interface and fast performance. 
Common complaints include limited color options and higher price compared to competitors

Input: 50 customer reviews
Output: "The product receives high marks for design and functionality. 
Customers frequently mention the intuitive interface and fast performance. 
Common complaints include limited color options and higher price compared to competitors

Meeting Notes Summary

Input: 4 weekly team meetings
Output: "Team focused on Q1 goals and resource allocation. Key decisions: 
hiring two new developers, postponing Feature X to Q2, implementing new CI/CD pipeline. 
Action items assigned for documentation updates and client presentations

Input: 4 weekly team meetings
Output: "Team focused on Q1 goals and resource allocation. Key decisions: 
hiring two new developers, postponing Feature X to Q2, implementing new CI/CD pipeline. 
Action items assigned for documentation updates and client presentations

Input: 4 weekly team meetings
Output: "Team focused on Q1 goals and resource allocation. Key decisions: 
hiring two new developers, postponing Feature X to Q2, implementing new CI/CD pipeline. 
Action items assigned for documentation updates and client presentations

Model Information

Model Used: Snowflake managed model
Context Window: 128,000 tokens per row (unlimited rows)
Max Output: 8,192 tokens

Limitations & Considerations

Input Processing

Each row can contain up to 128,000 tokens
Aggregates across all rows in group
No hard limit on number of rows

Cost

Billing based on input and output tokens
All text in aggregated column counts as input
Automatic prompt overhead added

Output Quality

Works best with coherent, related content
May produce generic summaries for disparate content
Quality improves with focused grouping

Regional Availability

AWS US West/East: ✓
Azure East US: ✓
EU regions: ✓
Cross-region inference: ✓

Best Practices

1. Group Data Appropriately

-- Good: Grouped by meaningful dimension
SELECT 
    product_category,
    price_tier,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM reviews
GROUP BY product_category, price_tier;

-- Less effective: Too broad
SELECT AI_SUMMARIZE_AGG(review_text) FROM

-- Good: Grouped by meaningful dimension
SELECT 
    product_category,
    price_tier,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM reviews
GROUP BY product_category, price_tier;

-- Less effective: Too broad
SELECT AI_SUMMARIZE_AGG(review_text) FROM

-- Good: Grouped by meaningful dimension
SELECT 
    product_category,
    price_tier,
    AI_SUMMARIZE_AGG(review_text) AS summary
FROM reviews
GROUP BY product_category, price_tier;

-- Less effective: Too broad
SELECT AI_SUMMARIZE_AGG(review_text) FROM

2. Filter Before Aggregating

SELECT 
    product_id,
    AI_SUMMARIZE_AGG(feedback_text) AS critical_feedback
FROM customer_feedback  
WHERE sentiment_score < -0.5  -- Only negative feedback
  AND feedback_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    product_id,
    AI_SUMMARIZE_AGG(feedback_text) AS critical_feedback
FROM customer_feedback  
WHERE sentiment_score < -0.5  -- Only negative feedback
  AND feedback_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    product_id,
    AI_SUMMARIZE_AGG(feedback_text) AS critical_feedback
FROM customer_feedback  
WHERE sentiment_score < -0.5  -- Only negative feedback
  AND feedback_date >= CURRENT_DATE - 30
GROUP BY

3. Combine with Metrics

SELECT 
    department,
    COUNT(*) AS ticket_count,
    AVG(resolution_time_hours) AS avg_resolution_time,
    AI_SUMMARIZE_AGG(ticket_description) AS common_issues
FROM support_tickets
WHERE created_date >= CURRENT_DATE - 7
GROUP BY

SELECT 
    department,
    COUNT(*) AS ticket_count,
    AVG(resolution_time_hours) AS avg_resolution_time,
    AI_SUMMARIZE_AGG(ticket_description) AS common_issues
FROM support_tickets
WHERE created_date >= CURRENT_DATE - 7
GROUP BY

SELECT 
    department,
    COUNT(*) AS ticket_count,
    AVG(resolution_time_hours) AS avg_resolution_time,
    AI_SUMMARIZE_AGG(ticket_description) AS common_issues
FROM support_tickets
WHERE created_date >= CURRENT_DATE - 7
GROUP BY

4. Use for Report Generation

CREATE TABLE weekly_summaries AS
SELECT 
    DATE_TRUNC('week', event_date) AS week,
    event_type,
    COUNT(*) AS event_count,
    AI_SUMMARIZE_AGG(event_details) AS weekly_summary
FROM events
WHERE event_date >= CURRENT_DATE - 90
GROUP BY week,

CREATE TABLE weekly_summaries AS
SELECT 
    DATE_TRUNC('week', event_date) AS week,
    event_type,
    COUNT(*) AS event_count,
    AI_SUMMARIZE_AGG(event_details) AS weekly_summary
FROM events
WHERE event_date >= CURRENT_DATE - 90
GROUP BY week,

CREATE TABLE weekly_summaries AS
SELECT 
    DATE_TRUNC('week', event_date) AS week,
    event_type,
    COUNT(*) AS event_count,
    AI_SUMMARIZE_AGG(event_details) AS weekly_summary
FROM events
WHERE event_date >= CURRENT_DATE - 90
GROUP BY week,

Common Use Cases

Executive Reporting

SELECT 
    'Q1 2024' AS quarter,
    AI_SUMMARIZE_AGG(project_update) AS quarterly_summary
FROM project_updates
WHERE update_date BETWEEN '2024-01-01' AND '2024-03-31'

SELECT 
    'Q1 2024' AS quarter,
    AI_SUMMARIZE_AGG(project_update) AS quarterly_summary
FROM project_updates
WHERE update_date BETWEEN '2024-01-01' AND '2024-03-31'

SELECT 
    'Q1 2024' AS quarter,
    AI_SUMMARIZE_AGG(project_update) AS quarterly_summary
FROM project_updates
WHERE update_date BETWEEN '2024-01-01' AND '2024-03-31'

Customer Insights

SELECT 
    customer_segment,
    COUNT(DISTINCT customer_id) AS customer_count,
    AI_SUMMARIZE_AGG(feedback) AS segment_insights
FROM customer_feedback
GROUP BY

SELECT 
    customer_segment,
    COUNT(DISTINCT customer_id) AS customer_count,
    AI_SUMMARIZE_AGG(feedback) AS segment_insights
FROM customer_feedback
GROUP BY

SELECT 
    customer_segment,
    COUNT(DISTINCT customer_id) AS customer_count,
    AI_SUMMARIZE_AGG(feedback) AS segment_insights
FROM customer_feedback
GROUP BY

Competitive Analysis

SELECT 
    competitor_name,
    AI_SUMMARIZE_AGG(mention_text) AS competitor_mentions
FROM social_media_monitoring
WHERE mention_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    competitor_name,
    AI_SUMMARIZE_AGG(mention_text) AS competitor_mentions
FROM social_media_monitoring
WHERE mention_date >= CURRENT_DATE - 30
GROUP BY

SELECT 
    competitor_name,
    AI_SUMMARIZE_AGG(mention_text) AS competitor_mentions
FROM social_media_monitoring
WHERE mention_date >= CURRENT_DATE - 30
GROUP BY

Related Functions

AI_AGG - For custom aggregation prompts
AI_COMPLETE - For detailed analysis
AI_SENTIMENT - For sentiment-focused aggregation

Interested to Learn More?
Try Out the Free 14-Days Trial

Start free trial

Learn

Feb 18, 2026

BigQuery Global Queries: How to Run Cross-Region SQL in 2026

Analytics

Feb 18, 2026

Context Engineering and AI Quality for Data Teams

Product

Feb 18, 2026

Accelerate Analytics Development with Paradime and Tableau

Product

Feb 18, 2026

Accelerate Analytics Development with Paradime and Tableau

Experience Analytics for the AI-Era

Start your 14-day trial today - it's free and no credit card needed

Start for free

Experience Analytics for the AI-Era

Start your 14-day trial today - it's free and no credit card needed

Start for free

Experience Analytics for the AI-Era

Start your 14-day trial today - it's free and no credit card needed

Start for free

Platform

Resources

ADD-ONs

Industries

About

Legal

Made with ❤️ in San Francisco ・ London

*dbt® and dbt Core® are federally registered trademarks of dbt Labs, Inc. in the United States and various jurisdictions around the world. Paradime is not a partner of dbt Labs. All rights therein are reserved to dbt Labs. Paradime is not a product or service of or endorsed by dbt Labs, Inc.

Start for free

Platform

Resources