We’re excited to share a major milestone in AI: Snowflake Intelligence is now generally available. Built on Snowflake Cortex and featuring new agentic capabilities, Snowflake Intelligence enables anyone — from business leaders to analysts — to explore and act on data simply by asking questions in natural language, without writing SQL or building dashboards by hand. 

Under the hood, Snowflake Intelligence plans and executes multi-step analytical workflows, connecting structured and unstructured data, applying models, and returning trusted insights, all while maintaining Snowflake’s enterprise-grade governance, security, and observability. The result is faster analysis, smarter automation, and a single, AI-native interface that makes your organization’s data accessible and actionable.

And behind all AI is great data. Fivetran is the leading platform for data integration, allowing you to automatically, reliably, and scalably integrate both structured and unstructured data into destinations such as Snowflake. Snowflake Intelligence builds on this robust data foundation with a comprehensive suite of AI tools in the Snowflake cloud.

How Snowflake Intelligence works

Snowflake Intelligence is based on several key capabilities:

• The orchestration of Cortex Agents, using a network of specialized AIs to coordinate reasoning, planning, and tool usage, as well as optimizations to make multi-step reasoning and large model use efficient
• Support for querying and retrieving structured and unstructured data through Cortex Analyst and Cortex Search, respectively
• Semantic models and knowledge indexing to map representations in data to real-world business concepts
• Governance, observability, and auditability to ensure that each reasoning step is visible and that data assets are used compliantly and accountably

Together, these capabilities enable you to converse with your data and even set up simple automations. You no longer need skills or experience using SQL, BI platforms, or ML notebooks to answer basic and pressing analytics questions.

How Fivetran and Snowflake Intelligence fit together

From raw data to conversational analytics, the architecture that combines Fivetran and Snowflake Intelligence looks like the following:

1. Fivetran extracts both structured and unstructured data from various sources and loads it into the Snowflake data platform.
2. With a semantic foundation, Cortex Analyst and Cortex Search can match structured and unstructured data to the real-world concepts they represent.
3. Snowflake Intelligence can be prompted by a user or triggered by an event. When prompted or triggered, it will activate a network of specialized agents that plan, coordinate, and reason to figure out the best response to your prompt, augmented by its semantic understanding of your data.
4. Fivetran Census routes data back into operational systems as needed, activating the data through operational analytics and business process automation.

Your first Fivetran and Snowflake Intelligence project

Getting started with Fivetran and Snowflake Intelligence is far easier than building reports and dashboards the conventional way. In short, the process consists of the following steps:

1. Use Fivetran to connect data from 2-3 high-value connectors (e.g., CRM, finance, operations) into Snowflake.
2. In Snowflake, define semantic views for the most important metrics through Cortex Analyst. 
3. Create a pilot agent tied to those views. Ask it some questions you know the answers to and see what comes out the other side. 
   
   1. Make it explain its reasoning.
   2. Take a look at its multi-hop reasoning chains to see how the answers were constructed.
4. As outputs improve in accuracy, continue iterating and adding more data sources and agents.
5. Use Snowflake Intelligence's automation capabilities to generate periodic reports and alerts.

Conversational analytics are a strong use case for agentic AI

Snowflake Intelligence promises to truly democratize the use of data by allowing users to conduct analysis in natural language. This capability is essential across all industries. The combination of Snowflake Intelligence with Fivetran, with its support for over 700 unique data sources and ready support for new, custom data connectors, is especially beneficial to organizations with data scattered across many sources and complex analytical workflows. 

Example use cases include:

• Retail/CPG – combining data from POS, inventory, and marketing to generative adaptive demand forecasting and granular sales optimization
• Finance – Conversational analytics for anomaly detection and explanation, as well as audit summaries
• Health/Life sciences – Combining patient outcomes, research, and claims data for analysis while ensuring regulatory compliance through auditability
• Energy and utilities – Detect operational inefficiencies, cost patterns
• Technology and software – Operational insights built from the full range of marketing, sales, product, and customer success data

From the standpoint of a typical user or team, Snowflake Intelligence radically simplifies the analytics workflow and enhances analytic agility, rapidly bridging the last mile between data and insight.

In the grander scheme of things, the capabilities already demonstrated by Snowflake Intelligence show that we are closer than ever to true mastery over data and its democratization, even as we are still in the early days of agentic AI. From conversational analytics, it’s only a few more steps to systems that can automatically recommend or even implement responses in real-time to changing conditions.

[CTA_MODULE]