What Dust does
Dust (dust.tt) is an enterprise “multiplayer” AI agent platform that connects AI models to a company’s internal knowledge and tools, and lets human teams and AI agents collaborate in shared workflows with permissions, auditability, and usage/cost visibility. The company positions Dust as an operating system for AI-driven enterprises—analogous to how Windows provided universal UI primitives—so that raw model capability becomes coordinated, compounding work across teams rather than isolated per-user chat. Core collaboration constructs include shared workspaces (Pods) where conversations, tasks, and files stay together as agent context. Dust also emphasizes model flexibility (multi-model choice), integration breadth via connectors/MCP, and governance features such as granular permissions, audit logs, and admin observability into agent interactions, adoption, and costs.
Dust targets “AI Operators” inside organizations—ops leads, support managers, RevOps/data analysts, growth/content teams, and other functional stakeholders—who want to run AI across departments while keeping humans in control of decisions, reviews, and handoffs. Business model appears to center on credit-based usage (Dust’s “credits” unit) with self-serve and enterprise tiers, including enterprise features like SSO, SCIM, pooled credits, audit logs, and data residency/single-tenant deployment options. Strategically, Dust’s messaging and recent funding indicate a focus on scaling adoption of connected, governed agent workflows inside enterprises rather than shipping generic chatbot UX, and it reports large rollout/usage at thousands of organizations.