AI Agents

Snowflake Introduces Always-On AI Agents That Work Across Enterprise Applications

AI agents are moving from responding to individual requests toward continuously executing business workflows.

On November 3, Snowflake announced new capabilities for its CoCo and CoWork products, including always-on bots designed to perform tasks in the background and involve employees when decisions or approvals are required. Snowflake says the bots will be available in private preview soon.

The announcement highlights a growing shift from conversational AI toward persistent enterprise automation.

Source: Snowflake — Snowflake CoCo and CoWork Power the Agentic Enterprise

From AI Assistants to Background Workers

Snowflake describes CoCo as an AI coding agent for data teams and CoWork as a personal work agent for knowledge workers. According to the company, CoCo has more than 9,100 accounts and CoWork more than 5,800.

The new bots act on a user's behalf, using that user's access rights and permissions, and can plan and complete multi-step goals such as preparing briefings for customer meetings or keeping data pipelines healthy.

These agents are designed to operate within tools teams already use. Snowflake for Slack and CoCo for Excel and Sheets are in public preview.

Recurring workflows can also be scheduled. Snowflake says each run of CoCo Automations keeps an audit trail and task history, and users can suspend or rerun jobs.

The distinction matters: traditional assistants wait for instructions, whereas persistent agents can monitor assigned responsibilities and initiate subsequent actions.

Business Context Becomes Essential

Snowflake is also expanding shared skills, memory, and reusable artifacts such as reports and dashboards.

These features are intended to help agents understand organizational workflows and retain useful context across repeated tasks. Users can view, update, or delete stored memories.

For enterprises, reliable automation requires more than language understanding.

An agent must access appropriate information, follow permissions, preserve workflow continuity, and produce results that employees can verify.

Snowflake positions its governed data infrastructure as a foundation for these requirements.

However, the announcement does not independently establish how frequently the new bots complete complex workflows without intervention.

The Agentic Age: Work Becomes Continuous

The broader economic significance is the possibility of shifting enterprise software from tools employees operate into systems that execute delegated responsibilities.

A financial analyst might receive regularly updated reports. A sales operations team could automate recurring data preparation. Developers could delegate routine data engineering tasks.

These are potential applications, not independently verified outcomes of the newly announced products.

The central challenge will be accountability.

Organizations must determine which actions agents may perform independently, which require approval, and how errors are detected and corrected.

Future competition among enterprise AI platforms may therefore depend less on conversational quality and more on measurable task completion, integration reliability, operating costs, and governance.

Snowflake's announcement illustrates that direction.

The next stage of enterprise AI will be defined not simply by agents that can answer questions, but by systems that can responsibly finish recurring work.

Sources

  1. Snowflake — Snowflake CoCo and CoWork Power the Agentic Enterprise, November 3, 2026

This article was prepared with AI assistance and reviewed by an Emerge Times editor before publication. Licensed CC BY 4.0: credit “Emerge Times (emergetimes.com)”. Usage & License. Featured images are AI-generated illustrations; any logos shown are trademarks of their owners, and Emerge Times is not affiliated with them.