Anthropic expands Claude Cowork agents to web and mobile

Claude Cowork moving from a desktop niche to a web‑and‑phone surface is a big moment for agentic AI. Instead of being tied to a single machine, Cowork can now run long‑lived workflows in Anthropic’s cloud—summarizing threads, drafting reports, reconciling data—while users supervise from anywhere. Anthropic’s own data underlines that these agents are mostly doing 'the work around the work': operational glue tasks and content prep, not hardcore coding. That’s exactly the mundane but ubiquitous labor slice AGI‑class systems are likely to eat first.([infoworld.com](https://www.infoworld.com/article/4194472/anthropic-expands-claude-cowork-to-web-and-mobile-as-enterprise-use-broadens.html))

Strategically, this pushes Anthropic deeper into the “always‑on co‑worker” category and away from being just another chat interface. For enterprises, the value is obvious: background agents that keep processes moving when people are offline. But it also raises governance stakes. Always‑connected agents touching email, docs, and ticketing systems expand the attack surface and make it easier for prompt‑injection or misconfiguration to have real impact. InfoWorld’s sources are already warning that deployment is outpacing security rigor.([infoworld.com](https://www.infoworld.com/article/4194472/anthropic-expands-claude-cowork-to-web-and-mobile-as-enterprise-use-broadens.html))

In the AGI race, whoever can safely operationalize these persistent agents at scale will gain massive real‑world behavioral data and customer lock‑in. Cowork’s expansion is Anthropic’s bid to secure that position before OpenAI, Microsoft, or others turn their own agents into standard operating infrastructure.