Anthropic

AI · Enterprise Software · Autonomous Agents

Launch

Anthropic launches Claude Managed Agents for autonomous enterprise AI agents

Most AI agent pilots fail because organizations can't operationalize them, not because the models fall short.

• Anthropic unveiled roughly 10 new enterprise capabilities for Claude, including Claude Managed Agents, spanning coding, cybersecurity, HR, and investment banking, ahead of a widely expected IPO.
• Claude Managed Agents gives businesses a software harness, memory system, and sandboxed environment so they can deploy fleets of autonomous Claude agents to run software projects unattended.
• A new integration with 1Password lets Claude agents authenticate with stored credentials for tasks like booking travel without ever exposing the underlying secret to the model or Anthropic's systems.
• The push builds on Claude Sonnet 5, which became the default model for Pro, Team, and Enterprise plans on June 30, 2026, and was tuned specifically for multi-step, tool-using agentic workflows.
• Anthropic is also backing Ode with Anthropic, a $1.5 billion AI implementation joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs built to help enterprises actually deploy the models.
• Gartner projects 40% of AI agent projects will be canceled by 2027, a 'delivery gap' Anthropic's APAC applied AI lead Natalie Mead cited at AWS Summit Taipei as the real barrier to adoption.
• The rollout signals frontier AI labs are now competing on deployment infrastructure as much as model quality, a shift that matters more as Anthropic heads toward a public listing.