Pulumi introduces Neo, an AI agent that can perform infrastructure tasks completely autonomously. The platform promises to reduce weeks of work to minutes, while maintaining enterprise governance.
Neo is supposed to differ from traditional automation tools in that it actually functions as a “teammate.” The agent understands dependencies, implements changes, monitors results, and keeps an eye on compliance at the same time. According to CEO Joe Duffy, “things that used to take weeks can now confidently be done in minutes.”
A crucial aspect of Neo is its built-in governance. The AI agent respects all existing team settings, including RBAC and policy-as-code configurations. Users can configure the level of autonomy, from fully protected to fully autonomous.
The agent operates with “progressive autonomy,” which involves human interaction with configurable automation levels. This means that teams can decide for themselves how much control they want to retain over critical infrastructure changes.
Practical results
Beta customers report consistent improvements: 10x more infrastructure delivery with existing teams, 75 percent faster deployment for operations that previously took weeks, and 90 percent fewer policy violations through automated governance.
“Pulumi Neo addresses our biggest challenge of eliminating the infrastructure bottleneck that slows down our entire engineering organization,” said Richard Genthner, Chief Information Security Officer at Boost Insurance.
Neo is also available through Pulumi’s Model Context Protocol (MCP) server, enabling integration with popular development tools such as Cursor, VS Code, and Claude Code. The platform is now available to all Pulumi users and customers, free of charge during the preview period.
Tip: Pulumi builds toolchains for Internal Developer Platform