Anthropic has introduced a new product named Claude Managed Agents, designed to streamline the development and operation of AI agents. This tool simplifies the process for developers by allowing them to define tasks, tools, and guardrails for building AI agents, while Claude Managed Agents takes care of the rest of the development cycle. The aim is to accelerate the transition of AI agents from prototype to launch within days instead of months.
The launch of Claude Managed Agents marks a significant step in the expansion of Anthropic’s enterprise business, which has been experiencing rapid growth. The company’s annualized recurring revenue stood at $9 billion by the end of 2025 and has now surged to $30 billion as of March 2026.
Most of Anthropic’s growth has been attributed to its Claude Platform, an enterprise solution that grants developers access to the company’s AI models via an API. With the introduction of Claude Managed Agents, the company anticipates further growth by relieving developers of the complexities of building AI agent infrastructure, allowing them to concentrate on enhancing user experiences.
Agents developed with Managed Agents will operate without the need for servers, security systems, or execution environments, as everything will be managed on Claude’s cloud infrastructure. The tool also features autonomous capabilities, enabling agents to independently call tools, manage context, and recover from errors.
Additionally, Managed Agents offers developers the ability to create long-running agents, manage other agents, and access APIs, databases, and internal tools. It provides enhanced monitoring control, allowing developers to track the tools used by the agent, its decision-making processes, and error locations.
Anthropic has made the tool available in public beta on the Claude Platform, starting immediately. During testing, the company observed a 10-point improvement in outcomes compared to standard prompting.
The launch of Claude Managed Agents follows Anthropic’s recent unveiling of its latest AI model, Claude Mythos Preview. The company claims that Claude Mythos has the capability to identify numerous cybersecurity vulnerabilities across applications and operating systems that were previously undetected by humans.
Anthropic is collaborating with over 40 tech firms to leverage Mythos in identifying and addressing vulnerabilities before potentially releasing the model to the public.