Dust, a French AI startup, aims to enhance team productivity by breaking down internal silos, revealing crucial knowledge, and providing tools to create custom internal apps.
The company utilizes large language models (LLMs) trained on internal company data, offering team members new capabilities. Unlike other AI startups, Dust doesn’t focus on creating new LLMs.
You May Also Like: Spotify Is Shutting Down Its Standalone Live Audio App Permanently
Instead, the company builds applications using existing LLMs developed by OpenAI, Cohere, AI21, and others.They developed a platform for designing and deploying LLM apps and subsequently concentrated on a specific use case—centralizing and indexing internal data for LLM utilization.
Dust integrates connectors that continuously fetch data from platforms like Notion, Slack, GitHub, and Google Drive. This data is indexed for semantic search queries. When a user engages with a Dust-powered app, the relevant internal data is located, used as context for an LLM, and an answer is provided.
For instance, imagine joining a company and working on a project initiated some time ago. If your company values transparent communication, finding information in existing internal data is crucial.
However, the internal knowledge base might be outdated or difficult to search, such as when discussions took place in archived Slack channels.
Dust goes beyond being a superior internal search tool by not just returning results. It can extract information from multiple data sources and present answers in a more useful manner. It functions as an internal ChatGPT and serves as the foundation for developing new internal tools.
Dust collaborates with design partners to explore various implementation and packaging options for the platform. They envision multiple products within the enterprise data and knowledge worker domain, supported by suitable models.
The startup acknowledges the challenges surrounding data retention, hallucination, and other LLM-related issues, and considers the possibility of creating its own LLM for data privacy purposes.
In a seed funding round, Dust raised $5.5 million (€5 million) with Sequoia leading the investment, joined by XYZ, GG1, Seedcamp, Connect, Motier Ventures, Tiny Supercomputer, AI Grant, and several prominent angel investors.
You May Also Like: Amazon Unveils AWS AppFabric To Simplify Connectivity With SaaS Apps
Dust anticipates that LLMs will significantly transform how companies operate. The value of a product like Dust is amplified in companies that embrace radical transparency, written communication, and autonomy, rather than information hoarding, endless meetings, and top-down management.
By harnessing the potential of LLMs and improving productivity, Dust offers a competitive edge to companies that adopt these values, unlocking untapped potential for knowledge workers.
