Notion Introduces External Agents With Claude and Cursor Integration.

Notion announced on LinkedIn that it is introducing External Agents, a new feature that connects AI tools like Claude and Cursor directly into its team collaboration platform. Teams can now assign tasks to these agents from shared boards, @-mention them in discussions, and watch them execute work alongside human teammates.

The update extends Notion's Developer Platform, which the company unveiled to give developers and agents new capabilities for building on top of Notion's infrastructure. External Agents represents a practical application of that platform, embedding agent-driven automation into the daily workflows teams already use for project management and documentation.

Unlike standalone AI tools that operate in isolation, External Agents live within Notion's existing collaborative environment. A team working on a product launch, for instance, can create a board with tasks for both human members and AI agents. Assigning a task to Claude or Cursor works the same way as assigning it to a colleague. The agent receives the context, executes the work, and updates the board when finished.

External Agents rely on Notion's Developer Platform, which provides APIs and infrastructure for custom integrations. The platform includes Workers for syncing external data sources, custom tools that agents can invoke, and declarative schemas that define how data flows between systems.

When a team @-mentions Claude in a Notion page or assigns a task to Cursor on a board, the agent accesses relevant context from the workspace, executes the assigned work using its native capabilities (whether that's code generation, analysis, or content creation), then writes results back into Notion. The integration is bidirectional: agents read from Notion to understand what's needed and write back to deliver completed work.

The Developer Platform also supports continuous syncs. Teams can set up Workers that pull data from external tools like Zendesk or Linear into Notion databases on a schedule, ensuring agents always have access to the latest information when they execute tasks. This addresses a common pain point in agent workflows: stale or disconnected data that requires manual updates.

Notion has been steadily expanding its AI capabilities over the past year. The company previously integrated Claude agents for collaborative task automation, and more recently added Cursor agents to task boards for full-cycle development. External Agents formalizes and broadens these integrations under a unified framework.

The move reflects a broader industry shift toward embedding AI agents in team software rather than treating them as standalone tools. Companies are investing in platforms that let agents operate within existing workflows, where they can access organizational context and deliver results without requiring users to switch between applications.

For teams already using Notion as a central workspace, External Agents removes friction from delegating work to AI. Instead of copying context into a separate AI tool, running a task, then manually updating project boards, the entire loop happens within Notion. This is particularly relevant for technical teams using Cursor for development work or knowledge workers relying on Claude for research and analysis.

The feature also positions Notion as a coordination layer for hybrid human-AI teams. As agents become more capable of handling complex, multi-step workflows, teams need infrastructure that treats them as first-class participants rather than external services. Notion's approach, embedding agents in boards and discussions alongside human teammates, suggests a model where delegation to AI becomes as routine as assigning a task to a coworker.

External Agents is available now through Notion's Developer Platform. Teams interested in setting up agent integrations can access documentation and tools at developers.notion.com.