Anthropic’s New Product Aims to Handle the Hard Part of Building AI Agents

Anthropic has launched Claude Managed Agents, a new enterprise product designed to remove the significant technical barriers to deploying AI agents at scale. The tool provides developers with an out-of-the-box "agent harness"—the critical software infrastructure, including memory systems and tool integrations, that wraps around an AI model to enable it to take autonomous actions. It also features a built-in sandboxed environment for secure project execution and allows agents to run autonomously in the cloud for hours. This move capitalizes on Anthropic's surging enterprise business, which now boasts over $30 billion in annualized recurring revenue, largely driven by its Claude Platform API.

The product directly addresses what Anthropic identifies as a major gap: the complex distributed-systems engineering required to move from a capable model to a reliable, scalable agent. Katelyn Lesse, head of engineering for the Claude Platform, notes this previously required dedicated engineering teams. Now, companies can redirect that talent. In a demo, productivity startup Notion showed how it uses Managed Agents to automate a lengthy client onboarding process within its app, with a manager monitoring agent activity via a Claude Platform dashboard. This release intensifies the race with OpenAI's Frontier platform to dominate the enterprise AI agent space, a sector some investors believe could disrupt traditional SaaS companies.