Notion to Launch Custom AI Agents Next Week
What happened
Notion is set to launch custom AI agents, allowing users to delegate repetitive tasks and automate workflows within the platform. The move comes as the company revealed that 50% of all Notion databases are now built by AI. Notion's CEO described the strategy as separating the platform from the product, empowering users to use AI as a co-creator.
Why it matters
- The new feature, part of a larger update called Notion 3.0, allows users to create "Custom Agents" that can work on autopilot to perform multi-step tasks. These agents can be scheduled or triggered to execute workflows, such as compiling daily user feedback reports or triaging IT requests. - Notion's AI agents can access data from both the web and integrated platforms like Slack, Zendesk, and Google Drive to complete tasks, while respecting the security permissions of those applications. For instance, an agent could be tasked to pull user feedback from multiple sources and generate a cited report with recommendations for future product updates. - CEO Ivan Zhao envisions these agents as a way to automate repetitive work, freeing up human workers for more strategic roles. Internally, Notion is already using over 700 AI agents alongside its employees to handle tasks like meeting notes and IT requests. - This launch is a key part of Notion's shift to an "AI-first" platform. The company is also experimenting with a usage-based pricing model for these agents, betting that selling outcomes rather than seats will significantly expand its market. - In early testing with companies like Ramp, a single custom agent answered 4,000 questions in a couple of weeks, saving an estimated 2,000 hours of work. The initial public launch will target three main use cases: answering repetitive questions, triaging tickets and tasks, and automatically generating status updates. - Competitors are also heavily investing in AI agents. Asana has "AI Teammates," ClickUp has "AI Super Agents," and Coda's AI can trigger automations and act on structured data. - Looking ahead, Notion is also developing a standalone AI chat application that will have the full context of a user's Notion workspace and will be able to manage email and calendars. The company also plans to release more developer-focused features, a shift for the historically non-developer-centric platform. - To build its AI-native talent, Notion has taken unconventional recruiting steps, including hiring programmers as young as 16.
Key numbers
- The move comes as the company revealed that 50% of all Notion databases are now built by AI.
- - The new feature, part of a larger update called Notion 3.0, allows users to create "Custom Agents" that can work on autopilot to perform multi-step tasks.
- Internally, Notion is already using over 700 AI agents alongside its employees to handle tasks like meeting notes and IT requests.
- In early testing with companies like Ramp, a single custom agent answered 4,000 questions in a couple of weeks, saving an estimated 2,000 hours of work.
What happens next
- These agents can be scheduled or triggered to execute workflows, such as compiling daily user feedback reports or triaging IT requests.
- For instance, an agent could be tasked to pull user feedback from multiple sources and generate a cited report with recommendations for future product updates.
- This launch is a key part of Notion's shift to an "AI-first" platform.
Quick answers
What happened in Notion to Launch Custom AI Agents Next Week?
Notion is set to launch custom AI agents, allowing users to delegate repetitive tasks and automate workflows within the platform. The move comes as the company revealed that 50% of all Notion databases are now built by AI. Notion's CEO described the strategy as separating the platform from the product, empowering users to use AI as a co-creator.
Why does Notion to Launch Custom AI Agents Next Week matter?
The new feature, part of a larger update called Notion 3.0, allows users to create "Custom Agents" that can work on autopilot to perform multi-step tasks. These agents can be scheduled or triggered to execute workflows, such as compiling daily user feedback reports or triaging IT requests. Notion's AI agents can access data from both the web and integrated platforms like Slack, Zendesk, and Google Drive to complete tasks, while respecting the security permissions of those applications. For instance, an agent could be tasked to pull user feedback from multiple sources and generate a cited report with recommendations for future product updates. CEO Ivan Zhao envisions these agents as a way to automate repetitive work, freeing up human workers for more strategic roles. Internally, Notion is already using over 700 AI agents alongside its employees to handle tasks like meeting notes and IT requests. This launch is a key part of Notion's shift to an "AI-first" platform. The company is also experimenting with a usage-based pricing model for these agents, betting that selling outcomes rather than seats will significantly expand its market. In early testing with companies like Ramp, a single custom agent answered 4,000 questions in a couple of weeks, saving an estimated 2,000 hours of work. The initial public launch will target three main use cases: answering repetitive questions, triaging tickets and tasks, and automatically generating status updates. Competitors are also heavily investing in AI agents. Asana has "AI Teammates," ClickUp has "AI Super Agents," and Coda's AI can trigger automations and act on structured data. Looking ahead, Notion is also developing a standalone AI chat application that will have the full context of a user's Notion workspace and will be able to manage email and calendars. The company also plans to release more developer-focused features, a shift for the historically non-developer-centric platform. To build its AI-native talent, Notion has taken unconventional recruiting steps, including hiring programmers as young as 16.
