Anthropic opens Claude Managed Agents in public beta

Anthropic is releasing Claude Managed Agents in public beta. According to the company’s description, the product lets enterprises run preconfigured Claude agents on infrastructure managed by Anthropic.

With infrastructure under Anthropic’s management, platform operations sit outside the enterprise’s scope. Deployment teams still configure agents to match work requirements, but they do not have to build and maintain the entire system running behind them. This arrangement suits organisations that want to put agents into use while limiting the resources devoted to infrastructure.

The preconfigured element points to an approach that packages some initial settings, rather than leaving each enterprise to assemble everything from scratch. For enterprises, the hard part usually lies in standardising processes and defining clearly what an agent is allowed to do, not only in the language model.

Public beta status indicates the product is open to a wider set of trial users, but is not yet a finished release. Enterprises considering a trial should first define the scope of work assigned to agents, how activity will be monitored, and what to do when an agent returns unexpected results.

Source: anthropic.com