Notion plans to shut down its Gmail client Notion Mail on September 22 and go “all in” on AI agents to run inboxes, saying 50%+ of users do not open the inbox
Last year, Notion expanded its productivity suite to include Notion Mail, an AI-powered email client.
Context & Ripple Effects
Notion’s mail effort moved from a 2024 preview of AI-assisted organization, drafting, scheduling, and follow-ups to a free Gmail client on macOS and the web in 2025. The product was positioned as an extension of Notion’s productivity suite rather than a standalone email platform.
The planned closure reverses that client-centric approach less than two years after its preview and follows Notion’s 2024 acquisition of Skiff, whose email service was also slated to close. It is another example of a productivity company treating email workflow as a feature layer rather than a durable destination.
First-order effects
- Notion Mail users must transition away from the client before its September 22 shutdown, with Gmail remaining the underlying email service for those users.
- Notion will redirect mail-product investment from interface features such as sorting and drafting toward AI agents intended to handle inbox work without requiring users to open the inbox.
Second-order effects
- The shift raises the bar for Gmail add-ons and alternative mail clients: competing products need to show that a dedicated inbox remains valuable when automation can triage, schedule, and follow up in the background.
- For Notion, email becomes less a separate surface to win and more a source of workflow context for its broader AI capabilities, potentially tightening the connection between messages, scheduling, and other productivity tasks.
Third-order effects
- If users increasingly delegate routine mail handling, the email-client market may concentrate around underlying platforms and automation layers, while standalone interfaces become harder to sustain.
- The durable differentiator in productivity software may shift from organizing information to reliably executing cross-app tasks on a user’s behalf; that makes trust, control, and error handling central constraints on agent adoption.
The trend: This is part of a broader move from AI-assisted productivity interfaces toward agentic systems designed to perform routine work in the background.