Snowflake builds new intelligence that goes beyond RAG to query and aggregate thousands of documents at once

Key Takeaways:

• Snowflake Intelligence addresses shortcomings of RAG-based document analysis.
• Agentic Document Analytics simultaneously processes thousands of documents.
• The platform unifies data across multiple sources, from Slack to SharePoint.
• Enterprises gain sub-second queries and cohesive data governance.
• AI insights become accessible to business users, without extra pipelines.

A New Approach to Enterprise AI

Snowflake unveiled its Snowflake Intelligence platform at the BUILD 2025 conference, framing it as a milestone in enterprise AI. By bringing structured and unstructured data together within a single governed environment, the company claims to resolve one of the biggest bottlenecks in current AI deployments: the inability to aggregate and analyze large volumes of nontraditional data quickly.

Understanding RAG’s Limitations

Traditional retrieval augmented generation (RAG) solutions excel at finding relevant text snippets but falter when organizations need to ask bigger questions—like counting references across tens of thousands of documents or summing numerical data trapped in PDFs. As Jeff Hollan, head of Cortex AI Agents at Snowflake, put it, “The pattern I think about with RAG is it’s like a librarian, you get a question and it tells you, ‘This book has the answer on this specific page.’” That approach can struggle when an enterprise must parse vast records in one go.

Inside Snowflake Intelligence

Agentic Document Analytics, a key addition to Snowflake Intelligence, solves this by making unstructured text as queryable as tabular data. This upgrade provides a structured view of thousands of documents at once, allowing users to run queries akin to “Show me a count of weekly mentions by product area in my customer support tickets for the last six months.” In practice, that means business-critical answers come from the same platform that handles transactional records and other structured data, limiting the need for separate pipelines or multiple databases.

Unifying Structured and Unstructured Data

What sets Snowflake’s approach apart is how it treats PDFs, Slack messages, Microsoft Teams data, and Salesforce records—as part of a single, integrated platform. Rather than forcing enterprise teams to wrestle with external vector databases or duplicative AI systems, Snowflake Intelligence parses, indexes, and analyzes text internally. All within a governed security boundary that satisfies tight compliance demands.

What This Means for Enterprises

For businesses, the ability to handle sophisticated queries across thousands of documents in sub-seconds opens new possibilities for support analysis, revenue forecasting, and competitive intelligence. It also democratizes AI analytics: Instead of specialists building custom data pipelines to handle unstructured data, any data-savvy user can run advanced queries and glean insights from content that was once scattered.

Charting the Path Forward

Snowflake executives see this development as a catalyst to get more organizations off the sidelines and into AI-driven innovation. “We have lots of organizations already getting value out of AI,” noted Christian Kleinerman, EVP of product at Snowflake. With Snowflake Intelligence, they aim to unify data and analytics under one umbrella, an approach that could shape the future of enterprise AI as more businesses move beyond basic text retrieval to comprehensive insight generation.