Who is best placed in modern enterprises to put artificial intelligence (AI) to work in the most effective way possible? The answer, surely, is the teams and departments that are going to be using that AI, rather than the IT function, say, or an external consultant who barely knows the business. Three year-old technology business Dust , which is today announcing a $40 million Series B financing round, wants to help more enterprises turn that idea into a reality.

“While there is widespread recognition of the potential of agentic AI, many companies are bringing in third parties to develop and implement their agents for their teams,” says Dust CEO Gabriel Hubert, who co-founded the San Francisco-based business in 2023 with Stanislas Polu. “We take the opposite approach, enabling teams themselves to design agents, building on the expertise they naturally have in their workflows and systems.”

The word “team” is particularly important to Dust. Lots of developers have built tools that enable individual employees – often tech-savvy members of staff – to develop AI agents that assist them in key tasks each day. Plenty of software now comes with copilot functionality that effectively enables individuals to use AI in this way, barely noticing that they’re doing so. The vibe coding phenomenon has enabled some employees to go ever further. However, this focus on the individual is destined to produce underwhelming results, Dust believes, since employees almost always work as part of a broader team, collaborating with colleagues in their own department and across the enterprise. When only one member of the team is leveraging AI – or each member is using their own agents – the impact is bound to be disappointing.

“The nature of work is too complicated for isolated single agents to be transformative,” argues Hubert. “Much of what we do isn’t contained in a self-contained task that an agent can take over and repeat over and again; rather people work with many different colleagues, often on open-ended projects that are never the same as the last one.”

Dust describes its solution to this challenge as “multi-player AI”. Its platform enables business teams to collectively build, deploy and manage AI agents that can take part in the complex ways in which groups of staff work together and with other groups. The platform features a shared collaboration space where teams and AI agents work together on day-to-day tasks; the platform integrates with the business’s existing software tools and its data sources – within privacy and security constraints – to maximize the impact of each new agent.

The company’s vision is for key members of staff inside functions such as operations, marketing and sales to use the platform to build and deploy AI agents without having to wait for a colleague from engineering, or an external consultant. Hubert says Dust is simple enough even for non-technical employees to use it. “We’ve tried to make it pretty low activation energy,” he says.

Users seem to agree. Since its launch, Dust has already signed up more than 3,000 organisations; in aggregate, these customers have collectively used the platform to launch more than 300,000 individual agents.

Damien Laborie, director of data and AI strategy at French IT firm Syxperiane, one of those customers, says solutions such as Dust are going to be essential for firms trying to exploit agentic AI. “Without writing code, without solving existential questions about data architecture and agent orchestration, it's increasingly ready to use,” he says. “The managers are the ones responsible for delivering results; once they're convinced AI can help them, they become our strongest ambassadors.”

These customer success stories have got investors excited. Today’s $40 million round is led by Abstract and Sequoia, with participation from Snowflake and Datadog. It takes the total amount of funding raised by the company above $60 million.

Konstantine Buhler, a partner at Sequoia, says Dust’s platform reflects the way businesses work in practice. “We’re in the early innings of a massive shift in how organizations use AI,” Buhler argues. “Most enterprise AI today is single-player: one person, one prompt, no compounding. Dust is building the multi-player system, where agents and humans share context and work together across the entire company. Zero churn and 70% weekly active usage tell you this isn’t experimental anymore. This is how enterprises will actually operate.”