New capabilities connect approved business definitions, centralised governance and recurring workflows, as Snowflake seeks to help enterprises deploy AI agents across data and business teams while managing access, costs and oversight within a shared platform.
Â
Snowflake has announced additions to its data and AI platform aimed at helping enterprises move AI agents from experimental projects into operational use. Unveiled at Snowflake Expedition 2026 on November 4, the updates cover business definitions, governance, cost management and autonomous workflows for data teams and knowledge workers.
The announcements include Business Ontology and Cortex AI Gateway within Snowflake Horizon Catalog, alongside new capabilities in Snowflake CoCo, its coding agent for data teams, and Snowflake CoWork, its personal work agent for knowledge workers. The products and features are at different stages of availability, ranging from general availability to private preview.
“The enterprises that win won't be the ones with the most agents. They'll be the ones whose agents know exactly what the business means by accessing the right data, abide by governance and guardrails specified, and can account for actions taken by agents,” said Christian Kleinerman, EVP of Product, Snowflake.
“That’s the difference between an agent that guesses and an agent the business can run on. Snowflake already holds the data, context, and governance enterprises rely on. By extending Snowflake Horizon Catalog with Business Ontology and Cortex AI Gateway, context and governance stop being afterthoughts and become the foundation every agent is built on for the agentic enterprise.”
Shared definitions for business data
Business Ontology, scheduled to enter public preview soon, is designed to give agents access to approved definitions of business concepts such as net revenue. Teams can assign an accountable owner to each definition and link it to underlying data inside and outside Snowflake.
The capability sits within Horizon Context and works with Semantic Views, which calculate defined metrics, and Cortex Sense, which infers context that has not already been defined. Snowflake said approved definitions take precedence, while logged changes allow organisations to trace how a business term was defined at a particular time.
The company cited research it sponsored with MIT Technology Review Insights, in which 55% of executives identified a lack of business context as the primary cause of low-quality AI agent insights. Fragmented or inaccessible data followed at 45%.
Conducted in July and August 2026, the research surveyed 400 senior data and technology executives globally at organisations with annual revenue of at least $500 million.
Centralised controls for agent activity
Cortex AI Gateway, now generally available, applies policies to agent connections across cloud environments, frameworks and AI providers. According to Snowflake, it extends existing access controls, authentication and permissions to agents interacting with enterprise data through Model Context Protocol servers.
Generally available capabilities include cost governance and observability, allowing teams to attribute spending, enforce limits and record agent activity within their Snowflake accounts.
Other controls remain in preview. Cortex AI Gateway Tools and Cortex AI Guardrails are in public preview, while Dynamic Model Routing, Agent Trace Intelligence and Agent Abuse Detection are in private preview. These capabilities are intended to support tool-call authorisation, protection against prompt injection, model selection and analysis of agent behaviour.
Automation for recurring work
CoCo Automations, now generally available, allows data teams to schedule recurring or long-running workflows, including pipeline checks and report preparation. Each run retains an audit trail and task history, with options to suspend or rerun jobs.
Snowflake also announced CoCo and CoWork Bots in private preview. Users can assign goals such as preparing customer meeting briefings or monitoring data pipelines, with the bots planning and executing work in the background. The company said the bots inherit users’ access rights and permissions and seek input when decisions or approvals are required.
Integrations through Snowflake for Slack and CoCo for Excel and Sheets are in public preview. Skills and Plugin Sharing is generally available, while Profiles in CoCo and CoWork are in public preview.
Enterprise use of AI agents
Shelter Mutual Insurance Company is among the organisations using Snowflake’s agent capabilities in operational workflows.
“It would take a lot of time to create semantic layers in the past, but things have changed drastically in the last three months since we started using Snowflake CoCo. Additionally, every piece of data flows live into Snowflake CoWork,” said Vibhor Gupta, Vice President of Enterprise Data & AI, Shelter Mutual Insurance Company.
“We now have access to real-time insights on demand directly in our operational processes. With eight AI agents in production and more underway, we’ve changed how data drives decisions from the claims desk to the C-suite.”
