Editorial illustration for Snowflake Introduces Cortex AI Gateway to Control Enterprise AI Agents
Snowflake Cortex AI Gateway Controls Enterprise Agents
Snowflake rolled out a batch of new AI governance tools at its Expedition 2026 conference on November 4, aimed squarely at a problem enterprises keep running into: AI agents that work fine in a demo but fall apart once they touch real business data at scale. The company's answer comes in two main pieces. Business Ontology, a new feature inside Snowflake Horizon Catalog, gives agents access to approved definitions of business terms and ties them directly to the underlying data, so an agent calculating "net revenue" uses the same definition a finance analyst would.
Cortex AI Gateway, also housed in Horizon Catalog, acts as a central control point for agents, models and tool calls, enforcing access policies and authentication across whatever AI provider or framework a company happens to be running. Snowflake also added automation features to CoCo, its coding agent for data teams, and to CoWork, pushing both products further toward autonomous execution rather than just assistance. The pitch from Snowflake's leadership is less about agent volume and more about whether those agents actually understand what the business means and operate inside the rules set for them.
Snowflake has introduced Business Ontology, Cortex AI Gateway and new capabilities in its CoCo and CoWork products to help enterprises deploy AI agents in production with business context, governance and autonomous execution.
Why this matters
Snowflake is betting that the next phase of enterprise AI isn't about better models, it's about plumbing: who gets to call which agent, under what policy, with what business definitions attached. That's a reasonable read of where things are stuck. Plenty of teams have agents that work in a demo and then stall the moment they need to touch production data governed by legal, security, and three different business units with conflicting definitions of "revenue."
For developers and founders building on top of Snowflake, Cortex AI Gateway is worth watching less as a feature and more as a signal: platform vendors want to own the control layer between agents and enterprise systems, not just the models. That's where lock-in and pricing power will live. Business Ontology is the other half of the bet, standardizing meaning so agents don't hallucinate definitions.
The open question Snowflake didn't answer in this announcement is portability. If your agent's permissions and business context live entirely inside Horizon Catalog, what happens when you want to run it somewhere else?
Related Reading
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
From Archives Happiest Minds Q2 EBITDA rises 18.1% YoY as revenue gains 12.5%
From Archives Three Early Claude Code Testers Leave Anthropic Within 10 Days
Common Questions Answered
What is the Cortex AI Gateway and how does it help enterprises deploy AI agents?
The Cortex AI Gateway is Snowflake's new governance tool designed to control enterprise AI agents in production environments. It addresses the critical problem of AI agents that function well in demos but fail when deployed at scale with real business data, providing policy-based access controls and governance mechanisms for agent execution.
How does Business Ontology improve AI agent performance with enterprise data?
Business Ontology, a feature within Snowflake Horizon Catalog, gives AI agents access to approved definitions of business terms and directly ties them to underlying data sources. This ensures agents use consistent, governance-approved terminology when performing calculations and accessing production data, eliminating confusion from conflicting business definitions across different departments.
What specific challenges do enterprises face when deploying AI agents to production according to Snowflake?
Enterprises struggle with AI agents that work in controlled demo environments but fail when accessing production data governed by legal, security, and compliance requirements. The core issue is managing conflicting business definitions across multiple business units and ensuring proper governance, policy enforcement, and data access controls at scale.
What products did Snowflake enhance alongside the Cortex AI Gateway announcement?
Snowflake introduced new capabilities in its CoCo and CoWork products as part of the Expedition 2026 conference announcements. These enhancements work alongside Business Ontology and the Cortex AI Gateway to provide comprehensive AI governance and autonomous execution capabilities for enterprise workflows.
Why does Snowflake believe enterprise AI success depends on governance rather than better models?
Snowflake's perspective is that the next phase of enterprise AI advancement focuses on infrastructure and governance 'plumbing' rather than model improvements. The company argues that controlling who can call which agent, under what policies, and with what business definitions attached is the key bottleneck preventing AI agents from moving from demos to production at scale.
Further Reading
- AI agents just doubled inside the enterprise. Confidence rose faster than control did. - TechCrunch
- Snowflake ups AI agent context, governance game - Constellation Research
- Snowflake Unveils Business Ontology and Cortex AI Gateway to Govern the Agentic Enterprise - Enterprise Times
- Datawiza Introduces Agent Gateway for Enterprise AI Governance - Datawiza
- Snowflake Unveils Enterprise AI Agents for 24/7 Autonomous Operations - SBS News
