Notion integrates GPT-4.1 and Claude 3.7 AI models into platform

Notion’s latest enterprise AI toolkit introduces a strategic integration of multiple large language models, combining OpenAI’s GPT-4.1 and Anthropic’s Claude 3.7 directly into its productivity platform. This move represents a significant competitive play in the enterprise productivity space, where model providers themselves are increasingly building similar features into their own platforms. By offering model switching capabilities alongside new AI-powered meeting tools and enterprise search functions, Notion is betting that unified workspace functionality will prove more valuable to businesses than subscribing to multiple specialized AI services.

The big picture: Notion has launched an all-in-one AI toolkit that embeds multiple leading LLMs directly into its workspace, allowing users to switch between models without leaving the platform.

• The new features include AI meeting notes, enterprise search capabilities, a research mode, and direct access to both GPT-4.1 and Claude 3.7 within the Notion environment.
• Early adopters of these features already include notable tech companies like OpenAI, Ramp, Vercel, and Harvey.

Behind the technology: Notion built its new features using a strategic mix of OpenAI and Anthropic models alongside its own fine-tuned AI systems.

• The company chose different models for different tasks, acknowledging that reasoning models like Claude 3.7 excel at thoughtful analysis but aren’t always ideal for quick productivity tasks like meeting transcription.
• According to Sarah Sachs, Notion AI Engineering Lead, the company fine-tuned the models using internal usage data and feedback from trusted testers to optimize for “Notion retrieval tasks.”

Key features: The new Notion AI for Work offering transcribes meetings, connects with platforms like Slack and Gmail for enterprise search, and offers a Research Mode for document drafting.

• The meeting functionality automatically tracks and transcribes calls when Notion is added to users’ calendars.
• Enterprise search functionality works across internal documents and connected third-party applications like Microsoft Teams, GitHub, Google Drive, and Sharepoint.
• Users can chat with either GPT-4.1 or Claude 3.7 directly within Notion and even create Notion templates from these conversations.

Competitive landscape: Despite embedding third-party AI models, Notion faces direct competition from the very model providers it’s partnering with.

• OpenAI’s Deep Research functionality, Google’s similar offering, and Anthropic’s internet search capabilities all compete with Notion’s research features.
• The meeting transcription and summarization space is already crowded with specialized AI services.

Why this matters: Notion’s strategy centers on consolidation and simplicity, offering enterprises a single platform with comprehensive AI capabilities rather than requiring multiple subscriptions.

• The approach addresses the growing fragmentation in the AI productivity space by providing one unified workspace with multiple AI capabilities.
• Business and Enterprise subscribers with the Notion AI add-on can access these new features immediately.