Snowflake Announces Expansion of Snowflake Intelligence and Cortex Code
Snowflake’s announcement about Snowflake Intelligence and Cortex Code is quite in line with the general trend among major software vendors to build agentic AI platforms with control planes. ServiceNow acquired data.world, and Salesforce bought Informatica exactly for the purpose of building platforms for AI agents in early 2025. Microsoft announced AI Foundry built on top of the IQ layer in November 2025. Teradata announced Enterprise AgentStack as their new platform for agentic AI hosting in January 2026. All these vendors are motivated by the same demand from the market: to move from Copilot-based generic chatbots to business purpose-specific AI agents that understand business logic and can interact with one another.
As always, the devil is in the details: what those platforms are composed of and how they offer to control AI agents. Most of the platforms are built the old-fashioned way: All the controls are coded. Snowflake speaks about reusable analytics through saving the whole solution and reusing complete modules or models. It means that common semantics are still buried inside database models and code. By comparison: ServiceNow plans to ensure common semantics through the ontology and knowledge graph built by data.world, Salesforce intends to reuse knowledge graph created by Informatica, and Teradata creates its own Knowledge Graph to provide common semantics for AI agents. Microsoft arguably went further than everyone else by creating a holistic IQ layer that connects all structured and unstructured sources via core business semantics implemented as ontology and the business context implemented as graphs. However, even Microsoft defines a specific agent’s behavior inside Python code at the moment.
Thus, it looks like Snowflake has caught up to the pack by integrating with other data systems (like Databricks and Postgres) and enabling building a broad AI ecosystem via MCP and agent communication protocol (ACP), but it has not outpaced their competition.
Our Take
In their announcement, Snowflake made the bold statement: “Snowflake Intelligence moves work forward for business users,” but we need to see the actual scope and scale of how much has been given to the business to manage.
Ideally, an agentic AI platform should have all the three major aspects declaratively implemented and controlled by the business (not code developers – who could be replaced by AI):
- Core business semantics, business context, and business rules.
- AI agent functional and behavioral role definition (i.e. what they are supposed and not supposed to do, and how they are supposed to interact with other agents and humans).
- Platform controls (external to the agents) over security, compliance with internal and external rules, and conflict resolution (something like AI platform police).