Dust

Dust.tt provides a horizontal, open-source platform for companies to design, deploy, and manage custom AI agents connected to their internal data and tools. Founded in 2022 by Stanislas Polu and Gabriel Hubert, the Paris-based company aims to enhance enterprise productivity by allowing non-technical users to build specialized AI assistants for specific workflows. The co-founders are serial entrepreneurs, having previously founded a data analytics company that was acquired by Stripe in 2015. Polu, who serves as CTO, is an alumnus of Polytechnique and Stanford and previously worked as a research engineer at OpenAI, focusing on the mathematical reasoning capabilities of large language models. Hubert is the CEO.

The company has evolved from a developer framework and browser extension to a comprehensive enterprise agent platform. Dust's core thesis is that the primary challenge in leveraging AI is not model capability, but the deployment and productization of these capabilities within organizations. The platform functions as an operating system for AI, providing the infrastructure to connect large language models (LLMs) to a company's unique data, including integrations with systems like Slack, Notion, Google Drive, and GitHub. This allows teams in various departments—such as sales, marketing, engineering, and customer support—to automate tasks, generate insights, and improve decision-making.

Dust takes a model-agnostic approach, supporting frontier models from providers like OpenAI, Anthropic, Google, and Mistral, allowing users to select the best model for their specific task. Its platform features a graphical user interface for creating prompt chains, a library of standard building blocks, and a proprietary scripting language for parsing model outputs. The company emphasizes security and privacy, with SOC 2 Type II certification and GDPR compliance, ensuring that client data is not used for model training. The business model is B2B SaaS, and the company has chosen a horizontal platform strategy over creating vertical-specific solutions, believing this allows for wider adoption and emergent use cases within organizations. Dust has raised a total of $21.5 million over two funding rounds, including a Series A led by Sequoia Capital.

Keywords: AI agent platform, large language models, enterprise AI, workflow automation, prompt engineering, LLM operations, data integration, AI assistants, model-agnostic, open-source AI, no-code AI builder, enterprise productivity, Stanislas Polu, Gabriel Hubert, Sequoia Capital, generative AI, AI orchestration, knowledge base integration, secure AI, custom AI agents