Dust hits $6M ARR helping enterprises build AI agents that actually do stuff instead of just talking

DustThe two-year-old artificial intelligence platform, which helps enterprises, reached the annual income to complete all the work, reached $ 6 million – only one year before $ 1 million. Acceptance of the company’s rapid growth, in a shift in the enterprise, which transforms into complex systems from a simple conversation to complex systems.

San Francisco-based startup, a new category of AI companies selected from Anthropin’s new category of anthropin’s ecosystem of anthropin’s Ecosystem on Thursday stressed a higher category from scratch to develop its AI systems.

“Users only want more than the negotiation interfaces,” he said. “Instead of creating a draft, they want to create an actual document automatically. They need the CRM notes that are updated without hand interference with a meeting summary.”

The dust platform goes beyond Chatbot-Style AI instruments that dominate the adoption of an early enterprise. Instead of just answering the questions, dust AI agents can automatically maintain Github Problems, calendar sessions, customer records and even maintaining code reviews – Industrial Security Protocols – enterprise-class security protocols.

How AI agents convert sales calls to automated GitHub tickets and CRM updates

The company’s approach is clear with a specific example of Hubert described: a business-business sales company using many powder agents to process transcripts of sales calls. An agent analyzes automatically automatically in Fighting Cactions with resonant and sales cards with prospects. At the same time, another agent determines the customer’s features, maps them on the road map, and in some cases create Github tickets for small features ready for development.

“Each call transcript will be analyzed by many agents,” he said. “You have a sales battle card Optimizer agent on sale, which is strong and prospectives, and resonates with the strong and prospect and intend to follow a process on the going and seller.”

This is active with automation levels Model Context Protocol (MCP)A new standard developed by an anthropic that allows EI systems to connect safely to external data sources and applications. Guillaume Princen, the head of Anthrop, described the MCP as “as a USB-C connector between AI models and applications”, which allows the agents to enter information, as a USB-C connector. “

Why the Klaude and MCP strengthen the next wave of AI automation

The success of the dust reflects the wider change in how enterprises approach the AI’s application. Companies such as powdered models, companies such as dust are increasingly capable of using basic models – especially anthropic’s Claude 4 Suite – and combine them with a special orchestral program.

“We just want to give our customers access to the best models,” Hubert said. “And I think it is now at the beginning of the leaders in anthropic, especially coding models.” The company serves a month to thousands of businesses from a small start to a small start with thousands of businesses and thousands of workers.

Anthropik’s cluster models saw a particularly strong adoption for coding tasks, and the company with the release of the latest Claude 4 models, the clod code increased by 300% increase in 300%. “Opus 4 is the strongest model for coding in the world,” he said. “We have already headed the coding race. We strengthen it.”

Enterprise security is obtained when AI agents can actually act

Switch to AI agents that can make real action throughout the work systems, provide new security complexity that are not available with simple chatbot applications. Dust, Hubert is called a “local permitting layer” separating the rights of information on the rights of data revenues.

“Permission and tool management, as well as information and tool management, AI agents are part of the planes to reduce the secretion of sensitive data in case of operating in many work systems,” the company explains technical documents. Agents are critical when creating GitHub problems, update CRM notes or change the documents within an organization’s technology stack.

The company implements an enterprise-grade infrastructure with zero information storage policies, which ensures that the sensitive job data developed by AI agents is not developed by the model provider. This creates a key concern for enterprises that intend to accept AI on a scale.

Rise of AI-local starting building in stock models instead of creating themselves

The growth of dust calls that anthropy cannot be regularly available without the opportunities of an ecosystem that arises from “AI’s native beginnings”. These firms develop their AI models, but are doing business by creating complex applications above the existing basic models.

“It has a very strong sense,” Princen said, “Princen said,” Princen said that the latest customers of these companies needed and this special use. ” “We provide tools for them to build tools and use their products to these special customers and use the cases they seek.”

This approach represents an important turn in the structure of the AI ​​industry. Instead of any company that needs to develop its AI opportunities, specialized platforms such as dust can provide the okestration layer that creates powerful AI models that are useful for special job applications.

$ 6 million in revenue signals for the company’s Software’s future

The success of companies such as powder shows that the company moves outside the practical exercise of the EI market outside the experimental stage. Instead of replacing the wholesale of human workers, these systems are designed to eliminate regular tasks and context between applications that allow employees to direct more valuable activities.

“By providing the universal AI primitives of the whole company more smarter, we also set the foundations for the relevant permission system, the future proven system,” he said.

The company’s customer base convinced that the AI ​​will change the operations of the work. “The common thread between all customers, cited against the future and it believes that this technology will change a lot,” he said.

AI models MCP becomes more skilled and protocols as adults, the difference between the information and actions will become the main differentiator in the enterprise market. Prompt income growth of dust shows that enterprises are willing to pay reward prices for AI systems that can only complement the real work instead of helping it.

The effects are extending to a wider structure of the enterprise program outside of individual companies. If AI agents can eliminate and automate workflows in connection work applications, the software may think about the purchase and workflow design – reduces complexity with preferential enterprise technology.

Perhaps the most open sign of this transformation naturally describes as Digital Workers, not like Hubert AI agents, but also to work every day. Companies such as APIS and integration platforms, powder, companies such as powders, prove that the future cannot require combining everything and cannot require the chaos we have already established.