An interview with Notion CEO Ivan Zhao on Custom Notion AI agents launching in the coming week, over 50% of Notion databases now being built by agents, and more
Context & Ripple Effects
Notion had already introduced customizable agents that can create documents and take background actions, extending its stated ambition to become an AI workspace that works for users. The new reported database-building share offers a concrete usage signal alongside that product direction.
The planned Custom Notion AI release moves agent use from a general feature set toward user-configured automation inside the workspace, where databases are a core organizing layer.
First-order effects
- Notion users will be able to configure agents for workspace-specific tasks, while the company gains a more direct route to make agent activity part of everyday database creation.
- With agents reportedly creating more than half of databases, Notion’s product experience and support burden increasingly depend on the quality, control, and reliability of agent-generated structures.
Second-order effects
- Workspace-software rivals face pressure to offer automation that acts on users’ behalf, not just AI assistance that responds within a document or chat window.
- Greater use of configured agents raises the value of permissions, review flows, and integrations: customers will need ways to govern what an agent can create or change across their workspaces.
Third-order effects
- If agent-created workspaces become normal, the competitive unit may shift from standalone productivity features to an agentic work surface that combines context, actions, and governance.
- The pattern favors products that can turn accumulated workspace data into supervised automation, though sustained adoption will depend on whether users trust agents with consequential organizational work.
The trend: This is part of the shift from AI as an in-app copilot to workflow-native agents that construct and operate the work environment itself.