Dust

Dust connects teams to AI assistants that answer questions, summarize documents, and automate tasks using company data from Slack, Notion, and Google Drive.

Description

Dust operates as a no-code agent builder that connects AI models to enterprise data repositories and internal tools through pre-built connectors spanning Slack, Google Drive, Notion, Confluence, and GitHub. Teams configure specialized agents through visual interfaces, enabling multi-tool workflows that combine semantic search, data analysis, and web navigation within unified workspaces rather than isolated chat sessions. The platform processes queries against connected knowledge bases while maintaining fine-grained access controls through Spaces—permission layers that segment sensitive information across organizational boundaries and user roles.

Customers

Clay, Doctolib, Qonto, Wakam, Alan, Malt, Payfit

What Problem Does Dust Solve?

Constantly switching between tools to find files, write content, or answer questions drags teams into busywork and kills productivity. This context-switching creates bottlenecks that slow down entire workflows. Dust lets teams build custom AI agents that automatically access all their company data and tools in one place, handling these repetitive tasks so employees can focus on higher-value work.

Pros

• Composable AI Agent Framework: Dust enables teams to build, test, and deploy AI agents tailored to business workflows using a modular, prompt-centric architecture.
• Multimodel Flexibility: Supports multiple foundation models (OpenAI, Anthropic, Mistral, Claude) in a single interface, allowing comparative experimentation and fallback strategies.
• Collaborative Agent Design: Offers a shared workspace for teams to co-develop agents, integrate tools, and manage knowledge grounding for enterprise alignment.

Cons

• Limited Prebuilt Agents: The platform emphasizes build-your-own flexibility, which may delay time-to-value for teams seeking turnkey solutions.
• Learning Curve for Complex Workflows: Designing agents with tool integration, vector stores, and orchestration may require technical onboarding and iteration.
• Scalability Considerations: Managing agents across teams and models can introduce complexity in versioning, usage tracking, and performance monitoring.

Investors

XYZ Venture Capital, Connect Ventures, Jean Charles Samuelian, Seedcamp, GG1, Tiny VC, Sequoia Capital, Motier Ventures, Nicolas Brusson

Last updated: September 8, 2025