Anthropic has officially rolled out its AI agent Claude Cowork to web and mobile platforms for Max subscribers. Previously, this tool was limited to the desktop app since its launch in January 2026. This move marks Anthropic's effort to transform what was once a technical agent into a versatile administrative assistant that works seamlessly across all devices for office workers.
Detailed Developments
According to TechCrunch, the new update allows users to start tasks on a computer, get status updates on their phone, and retrieve the final output at any time. Notably, Claude Cowork can run autonomously in the background even after the user closes their laptop or goes offline. When encountering decisions that require human judgment, the system pings the user's phone for approval before execution. To encourage this new wave of adoption, Anthropic is also extending doubled Cowork usage limits through August 5.
Background & Causes
Alongside this release, Anthropic published usage data from 1.2 million anonymized Claude Cowork sessions across more than 600,000 organizations in May. The results revealed an unexpected reality: the vast majority of users are not using the tool for coding. Specifically, business process operations (such as spreadsheet reconciliation and building reports) claimed the largest share at 33.4%. Content creation and copywriting came in second at 16.4%, while software development accounted for just 8.7% of total usage.
Technical & Technology Analysis
Technically, Claude Cowork is powered by the company's recently released Claude Sonnet 5 large language model. Unlike the desktop version, which accesses local files and the client's browser, the web and mobile versions process asynchronous tasks server-side. This architecture enables complex autonomous scheduled tasks, such as automatically parsing email threads, synthesizing meeting transcripts, and drafting follow-up emails at specific hours.
Expert Opinions & Insights
Moving task processing to the cloud has raised mixed opinions regarding security. According to a report by SiliconANGLE citing research from security firm Armadin, the desktop version of Cowork was previously found to have a critical sandbox escape vulnerability on Windows. Although Anthropic dismissed the risk because exploiting it requires pre-existing local access, security experts still warn that letting AI agents automatically process sensitive emails and documents opens up a new attack surface for prompt injection techniques.
Impact & Future
Bringing Claude Cowork to multiple platforms demonstrates Anthropic's ambition to dominate the workspace of global knowledge workers, directly competing with similar solutions from OpenAI. By framing AI as a tool to handle "the work around the work," Anthropic is trying to prove the tangible value of autonomous agents. For tech-savvy professionals in Vietnam, this presents a significant opportunity to streamline daily administrative workflows using always-on AI agents.