Dust

Just over a year after Dust was born, we’re excited to announce: - our series A round led by Sequoia Capital. - a new generation of multi-tool assistants for all users. - passing the $1M ARR mark with over 5k active subscribers to Dust. We’ll accelerate investment to get work to work better, abiding by our “𝗼𝗻𝗲 𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺 𝗳𝗼𝗿 𝗮𝗹𝗹, 𝗯𝘂𝘁 𝗻𝗼𝘁 𝗼𝗻𝗲-𝘀𝗶𝘇𝗲-𝗳𝗶𝘁𝘀-𝗮𝗹𝗹” vision. We believe in highly focused assistants that tackle specialized tasks and in the value of a unified platform to manage adoption, costs, and security. ✍️ Blog → https://lnkd.in/eNqzdHmQ 📰 More from: Techcrunch - https://lnkd.in/eRpgerPs Les Echos - https://lnkd.in/eNY7EkdU Bloomberg - https://lnkd.in/eHrczRRw As announced to users already, we’ve worked hard to make Dust assistants smarter. They can now handle complex workflows, thinking in multiple steps and using various tools to answer queries. ✍️ Blog → https://lnkd.in/e4bNM3z9 We’re proud to be partnering with some of the fastest-moving teams in the world, and excited to accelerate. PS — we’re hiring! Explore our careers page and reach out if you think you’re a fit for any of our open positions! https://lnkd.in/erc6_wsG

🔬 Dust Session for Builders: Unveiling Multi-Tool AI Assistants To Our Community Recently, we held our second Dust Session for Builders. These sessions are our way of sharing new developments and exploring ideas with the people who use our tools. This time, we focused on our latest feature: multi-tool AI assistants. Let us break it down for you: The core idea: An AI assistant that can use multiple tools to solve problems, much like how we might use different tools when tackling a complex task. These tools include: • Searching through data • Querying tables • Fetching recent information • Processing large datasets • Browsing the web • Running custom-built apps For instance, to compile a weekly report, the assistant might: 1️⃣ Retrieve recent messages from a Slack channel 2️⃣ Extract key information from those messages 3️⃣ Organize the data into a clear format 4️⃣ Send out a Slack highlight to the team As we continue developing these multi-tool assistants, we're grappling with questions like: • How do we make these tools even more useful and easy to use? • What new capabilities should we add? • How do we help builders develop more advanced use cases? We don't have all the answers yet. That's why these sessions are valuable - they allow us to think out loud and explore ideas together. ✍️ Dive deeper into Dust's multi-tool assistants and watch the replay: https://lnkd.in/dMrtjJWs What kind of problems do you think AI could help solve in your work? We're genuinely curious to hear your thoughts and ideas.

⭐️ Enhancing Legal Workflows: Didomi's CPO Leverages Dust AI assistants. Thomas Adhumeau, Chief Privacy Officer at Didomi, sought innovative ways to optimize his team's efficiency in handling complex privacy compliance tasks. By integrating Dust AI assistants into their workflow, they achieved: • 50% time savings on routine tasks. • Streamlined multilingual compliance checks. • Improved team knowledge-sharing and collaboration. The assistants, built with Dust, now support Q&A processes, legal research, and GDPR compliance checks across multiple jurisdictions, empowering the entire team to work more effectively. Thomas's insight: "𝐴𝐼 𝑖𝑠𝑛'𝑡 𝑠𝑒𝑝𝑎𝑟𝑎𝑡𝑒 𝑓𝑟𝑜𝑚 𝑦𝑜𝑢𝑟 ��𝑟𝑎𝑖𝑛; 𝑖𝑡 𝑤𝑜𝑟𝑘𝑠 𝑖𝑛 𝑐𝑜𝑛𝑗𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑖𝑡. 𝑅𝑒𝑔𝑢𝑙𝑎𝑟 𝑢𝑠𝑒 ℎ𝑒𝑙𝑝𝑠 𝑦𝑜𝑢 𝑙𝑒𝑎𝑟𝑛 𝑡𝑜 𝑐𝑜𝑚𝑚𝑢𝑛𝑖𝑐𝑎𝑡𝑒 𝑐𝑙𝑒𝑎𝑟 𝑖𝑛𝑠𝑡𝑟𝑢𝑐𝑡𝑖𝑜𝑛𝑠, 𝑒𝑛ℎ𝑎𝑛𝑐𝑖𝑛𝑔 𝑦𝑜𝑢𝑟 𝑝𝑟𝑜𝑏𝑙𝑒𝑚-𝑠𝑜𝑙𝑣𝑖𝑛𝑔 𝑐𝑎𝑝𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠." ✍️ Read the full story here: https://lnkd.in/eSvxfZTS

Bravo à Dust pour sa Série A de 16M$ menée par Sequoia 🔥 Les metrics sont solides : • 1M$ d'ARR en 1 an • un ratio de 70% weekly active users / monthly active users 👉 preuve d'une habitude sur un usage chaque semaine • 1 à 2 heures gagnées par employé par semaine sur des tâches variées • 80% des employés de certains de leurs clients utilisent des assistants IA chaque semaine Encore une bonne nouvelle pour l'écosystème IA français qui continue sa croissance 🇫🇷 L'occasion de redécouvrir notre interview de Gabriel Hubert, cofondateur et CEO 👇 (lien en commentaire)

Thibault Martin from our Business Operations team shares how he's using Dust to build AI assistants that make our data workflows 50% faster. With assistants like "analyticsGod" to query data via natural language and "sqlGod" to collaboratively write SQL, Thibault and the team: - Get instant answers to data questions without manually hunting for information; - Enable self-serve analytics for members of the team; - Spend less time on basic data tasks and more time on high-impact analysis. "Start simple and be iterative in improving your assistants. Even if the outputs aren't perfect at first, they'll save you a ton of time," Thibault advises. https://lnkd.in/edPu8Qe2

At PayFit, custom AI assistants are transforming how teams work ⤵️ - Meeting recaps generated automatically, keeping everyone aligned; - Content drafts created more quickly, allowing more time to refine the final version; - Job postings enhanced with custom assistants' feedback; - ... It began with the foresight of Ghislain de Fontenay Payfit's founder and Sylvain Grande Payfit's CPO. Alexandre Pineau then played a key role, using Dust to develop custom AI assistants by collaborating with employees across the company because they have the most profound understanding of their daily work and customers. ✍️ Blog: https://lnkd.in/e5J24kyk

🤖Building an AI Assistant Using LinkedIn and Dust. This comprehensive guide details leveraging your team's LinkedIn or CRM data to forge powerful connections using Dust. You can expect insights into crafting custom AI assistants that effectively map and use your team's professional networks. ✍️ Blog: https://lnkd.in/eN-iW-jQ

Have you ever wanted to extract structured data from unstructured sources? Slack, emails, tickets, documents... a ton of information gets lost after a while because it's not structured. We just released a new way to extract data from any data source and feed it to one of your custom assistants for processing in a few straightforward steps! Compile user feedback, extract bug reports, extract specific information from emails, gauge customers' sentiments... The possibilities are endless! The good thing is that we have a full working demo at your fingertips right here 👉 https://lnkd.in/gPV9EQTE

One of the most frequent questions we get from our users is: "Which LLM should I choose for my custom assistant?" At Dust, we've extensively tested models like Claude Opus, GPT-4o, Gemini Pro 1.5, and Mistral Large to understand their strengths and quirks. Technical specifications tell a part of the story. To find the ideal model, you need to experiment - immerse yourself in prompting the different models and observe how they tackle your specific challenges. That's the core philosophy behind Dust - enabling you to test and compare models on real tasks seamlessly. By iterating between them, you can accelerate the discovery process and find the model that aligns with your problem-solving approach. Once you've found your match, dedicate effort to mastering the nuances of interacting with —or prompting— the model for your specific requirements and needs. The payoff is a powerful assistant tailored to your unique context and reading/writing style. As the AI landscape evolves, one principle remains constant and encoded in our vision at Dust—to unlock these models' full potential; we want to help everyone invest in hands-on learning and customization. So dive in, experiment boldly, and find the models and assistants to help you push the boundaries of what's possible. Read more 👉 https://lnkd.in/eWgR3nkV

We're rolling out a suite of new features and updates that significantly enhance your experience with Dust. Here's a quick rundown of what's new: 1️⃣ 𝗚𝗣𝗧-𝟰𝗼 𝗶𝘀 𝗛𝗲𝗿𝗲: OpenAI’s latest model, GPT4-o, is now available on Dust! 2️⃣ 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁 𝗧𝗲𝗺𝗽𝗹𝗮𝘁𝗲𝘀 𝗚𝗮𝗹𝗹𝗲𝗿𝘆: Enjoy pre-filled instructions to ease the creation of your custom assistants. 3️⃣ 𝗣𝗿𝗼𝗺𝗽𝘁 𝗦𝘂𝗴𝗴𝗲𝘀𝘁𝗶𝗼𝗻𝘀: Hone your prompting skills thanks to automated suggestions to optimize your assistant's instructions. 4️⃣ 𝗠𝘂𝗹𝘁𝗶-𝗙𝗶𝗹𝗲 𝗨𝗽𝗹𝗼𝗮𝗱𝘀: Send several files at once directly through the chat bar. 🔒 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 𝗦𝗲𝗰𝗿𝗲𝘁𝘀: Safely store and utilize API keys, tokens, and passwords within Dust Apps. 🤖 𝗜𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗲 𝘄𝗶𝘁𝗵 𝗦𝗹𝗮𝗰𝗸: Automate tasks by integrating Dust assistants into your Slack workflows, making your day-to-day operations more streamlined. 🔧 𝗖𝘂𝘀𝘁𝗼𝗺𝗶𝘇𝗮𝗯𝗹𝗲 𝗜𝗻𝘁𝗲𝗿𝗰𝗼𝗺 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗼𝗻𝘀: Gain more control over your synced data with customizable settings for Intercom connections. 🔄 𝗥𝗲𝗳𝗿𝗲𝘀𝗵𝗲𝗱 𝗪𝗲𝗯𝘀𝗶𝘁𝗲: Our new website is out! 🔬 𝗕𝗲𝘁𝗮 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀: Try out our beta features like the Process Data action and Google Meet Transcripts summarization to manage large data sets and automate transcript handling. Check out the full details on our blog: https://lnkd.in/ejRz2J7F PS: Alban Dumouilla and Clement Bruneau will be at Vivatech on Friday. Feel free to come and say hi 👋