Notion just turned its workspace into a hub for AI agents

Notion is embarking on a new journey to redefine productivity software. In a recent live-streamed announcement, the company, renowned for its collaborative note-taking tool, unveiled an innovative developer platform aimed at enhancing the functionality of its custom AI agents. This platform is designed to foster connections with external agents and enable teams to construct automated multi-step workflows that can integrate data from any database. By introducing an orchestration layer, Notion is evolving from a simple note-taking application into a central hub where users and AI agents can collaborate seamlessly across various tools and data sources. Back in February, Notion launched its Custom Agents—AI assistants that handle mundane tasks, such as addressing frequently asked questions and automating workflows. Since then, over one million agents have been created by Notion users. However, these agents faced certain limitations as they could not interface with external data or utilize custom logic. Teams often resorted to using third-party automation tools or developing their own scripts to overcome these hurdles. "Historically, Notion hasn't been the most developer-centric platform," acknowledged Ivan Zhao, Notion's co-founder and CEO, during the livestream. "But we are evolving." The new platform allows teams to roll out their custom code through Notion’s cloud-based environment, known as Workers. This feature enables users to write their own logic and deploy it within a secure sandbox, ensuring it operates independently from other systems. It opens up possibilities for syncing data into Notion, crafting custom tools, and triggering actions through automated signals from other applications, without needing external infrastructure. Interestingly, users can even rely on their favored AI coding agents to write the code for them. Notion is making it easy for developers to explore these features by offering free access to the Workers until August. Additionally, the platform now supports syncing external data sources through an API, allowing users to pull data from systems like Salesforce and Zendesk, keeping their Notion databases updated. Zhao pointed out that this enhancement means users can utilize their Notion database as a versatile canvas for managing workflows and AI agents. Workers can also facilitate the creation of agent tools equipped with custom logic for more complex interactions with third-party systems via the Model Context Protocol (MCP), a new standard for connecting AI tools to external data. Moreover, users can now engage directly with external AI agents, assigning tasks and monitoring their progress as if they were integrated with Notion’s own custom agents. At launch, Notion supports several partner agents, including Claude Code, Cursor, Codex, and Decagon, with plans to introduce more in the future. To interact with the Developer Platform, developers will use the Notion CLI, a command-line interface available to Business and Enterprise Plan subscribers. This shift in strategy positions Notion as a more programmable platform, moving beyond just an application to a comprehensive solution for workflow automation. As businesses increasingly seek to automate knowledge work and develop internal AI systems, a platform that integrates agents, custom code, and real-time data is becoming essential. Zhao encapsulated the vision for the Notion Developer Platform, stating, "Any data, any tool, any agent—that’s the big picture."