Dust

92% of an entire consulting agency now uses Dust daily. Most enterprise AI projects struggle to hit 30% adoption. This was all because the people doing the work were the ones leading the transformation. Traditional AI rollouts follow the same (not very successful) pattern: leadership decides, IT implements, employees resist. Insign, a 20-year-old French consulting agency, inverted this. They created "The Transformers," a cross-functional group that experimented with tools before any mandates came down. Very deliberate. Put AI tools in people's hands, let them play, then build processes around what works. Within a year, they'd deployed 42 specialized agents across their entire delivery pipeline: 1/ A Tender Expert that analyzes 500-page documents and flags inconsistencies instantly → 30% faster response times 2/ A Smart Planner that generates complete Gantt charts in minutes, accounting for team availability and holidays → 50% time savings 3/ A Quality Guardian that verifies thousands of brand assets automatically, catching errors humans inevitably miss at scale Vincent Vitre, Insign's COO, puts it clearly: "AI is nothing without HI and EI: Human Intelligence and Emotional Intelligence." This is a message we, at Dust, like to reiterate for anyone deploying AI: adoption never fails because the technology isn't “ready.” It does because the implementation starts at the wrong end. When the people doing the work drive the transformation, they build systems they'll actually use. Bottom-up beats top-down when you're changing how teams think, not just what tools they have. Thank you Insign for pushing the boundaries of what's possible with Dust and serving as a model for successful AI deployment!

90% of Watershed’s team now uses Dust agents in their day-to-day work. Enterprise AI adoption typically stalls after the pilot phase. Teams run a few experiments, see promising results, then hit the wall of organizational inertia. Watershed took a radically different approach. Instead of building AI strategy from the top down, they amplified what was already working from the bottom up. Rather than teaching people about AI through demos and workshops, they empowered teams to build solutions for their actual work challenges. They embedded AI champions on every team to identify real problems AI could solve today. These weren't external consultants or dedicated AI specialists. They were the builders already on their teams: engineers who'd been experimenting since GPT-3, sales ops people who hack workflows together, data scientists who spot patterns others miss. Within months, AI transformed from experimental to essential across Watershed: 1/ Sales development reps built prospect research agents that follow their specific playbook 2/ Engineering teams use AI for design documentation completeness checks 3/ A third of the company relies on an AI performance review coach they developed internally 4/ Sales operations automated call note processing and Salesforce updates "Employees went from thinking AI was a fun novelty you could use for 1% of your job to understanding that, for many roles, if you're not using AI actively, you're not working as effectively as you could be.”

🎉 Engineering Gala, Closing Night of dotAI by Dust! After a full day of talks, debates, and deep dives into the world of AI, Dust will be wrapping up dotAI with the official closing party: the Engineering Gala. Join us for a night where the community comes together one last time for great music, amazing food, and even better company. Expect: - A relaxed, social evening with the brightest minds in AI - Drinks, bites, and late-night conversations - Live DJ to keep the energy going - A few surprises and hands-on activities (because engineers can’t *not* build things) No talks, no slides! Just good vibes and the chance to connect with the people behind the projects you’ve been talking about all day. Whether you’re here to celebrate, network, or just grab one last drink with your fellow builders, this is the way to wrap up dotAI. 📍Paris, France  📅 When: Nov. 6  ⚠️ Spots are limited  🎟️ RSVP now and join the closing chapter of dotAI.

Today we're introducing Frames: interactive data visualization built directly into Dust agents. Most business reports are one-way streets. You build them. Share them. Then spend the next week answering "Can we see this filtered by X?" or "What does this look like if we exclude Y?" Your agents now create customized content your team can explore and share, without coming back to you for every variation. How teams are already using Frames: → GTM teams share pipeline charts prospects can interact with during calls → Product teams build dashboards combining NPS, support tickets, and usage data → Engineering creates sprint retrospectives stakeholders can drill into → Operations transforms documentation into interactive workflows Built for collaboration, secure by default: → Share internally with workspace links → Share externally with secure, token-gated links → Control view → Works with your existing platforms (Slack, Notion, GitHub, and more) Adding Frames to your agents takes seconds. Open Agent Builder, select "Create Frames," and you're done. Available now for all Dust customers. See Frames in action in the comments with our H1 2025 growth report (yes, real numbers 😉) Click around, explore the data. This is what your agents can create. Not using Dust yet? Visit Dust to see how Frames can transform how your team works with data.

Really grateful to be on LinkedIn’s Top 20 Startups in France 2025. The real win is seeing our customers turn AI into outcomes across so many verticals and into use cases we never imagined when we started building Dust. At Dust, we focus on making AI helpful where work actually happens. Credit to the customers and champions who keep pushing for real change. If you’re leading AI transformation in your org, or if you want to integrate AI but don’t know where to start, reach out.

We're hosting TWO events this week: - Oct. 15: Engineering Night in Berlin 🇩🇪 Audience: software engineers -- speakers incl. Kirstin Taufertshöfer from Qdrant & our own Jules Belveze - Oct. 16: Behind the Agents in Paris 🇫🇷 Audience: anyone curious to see how Eria (Tifany Clemenceau & Melvin Duveau 💫) automated Sowefund's customer support -- speakers incl. Florian Gauthey Links to register in comments 👇

Wakam built an AI-powered legal department with Dust and cut legal contract analysis time by 50%. Their legal team was spending days extracting deadlines from contracts across 32 countries. Hours hunting through scattered systems for the right clause. More hours translating between English and French without missing compliance details. Because of this, the strategic work that moves the business forward kept getting pushed aside. Wakam spent 6 months trying to build their own AI solution, but it wasn't moving fast enough. They switched to Dust and built their first 3 AI agents. One agent reads, writes, and translates contracts while preserving every legal nuance. Another pulls critical data from their entire contract database: deadlines, pricing, and SLAs. A third monitors regulatory changes across multiple jurisdictions. Now: 1/ Contract analysis time is cut in half 2/ Document searches are 40-60% faster 3/ Their legal team can finally focus on strategy (and is much happier!) Today, Wakam runs 130+ AI agents across their entire company. But above all, their team is doing what they're best at: complex negotiations, strategic decisions, creative problem-solving. It's a common theme among our customers: the fastest transformation happens when you stop trying to build solutions from scratch for AI when something like Dust exists.

Drum roll please 🥁🥁🥁 issue #7 of our newsletter is out: 'Another one bites on Dust' 👑 (link in comments) - meet us IRL this week in London 🇬🇧 (👋 Alexandre Moog Christian Ryan Vincent Moens Anna Logacheva Amelie Deltombe Adèle Dugré Sophie Bergeret Faateh Dhillon Jules Belveze Kevin Straszburger) - news (incl. Gabriel Hubert interview with Diamond Bishop & Shah Ahmed) - strong use cases for sales (💪 Edouard Villette) and some easy wins AI-first GTM teams have already implemented (🙏 Gina Kabasakalis) - our little 'french touch' corner with even more use cases (✌️ Flavie Prevot Niels Rolland Yssine Matola Lobna Calleja Ben Hassine Tifany Clemenceau Melvin Duveau 💫) Onwards 🚀

London calling 🎸 We got THREE events next week in the Swinging City. Find yours and join us. It'll be fun and you'll learn a few things (and you'll get to meet a few of us IRL) :) - Oct. 6: Build your first AI Agent. Audience: anyone interested in learning about agents -- hosting this with Alexandre Moog from Paatch - Oct. 7: Engineering Night with technical content. Audience: software engineers -- speakers incl. Christian Ryan from Anthropic, Vincent Moens from Periodic Labs & our own Jules Belveze - Oct. 8: AI Marketing Night. Audience: AI-curious and AI-enthusiastic marketers -- speakers incl. Anna Logacheva from Spicy Lemon & our own Amelie Deltombe Links to register in comments 👇

In this talk, Gabriel, cofounder of Dust, breaks down 3 key insights from the Shopify memo that are driving the future of work: ✅ AI won't just change tools - it will change expectations of how people shape those tools ✅ Everyone (not just IT) should define how AI behaves ✅ The timeline for this transformation is NOW