Anthropic is fundamentally reshaping the accessibility and functionality of its AI agentic tool, Claude Cowork, by transitioning it from a desktop-bound application to a fully cloud-based, web, and mobile experience. This significant evolution, announced Tuesday, removes the previous requirement for a user’s computer to remain active for Cowork to perform its tasks, including scheduled operations. This move promises to enhance user convenience and broaden the appeal of agentic AI for knowledge workers by offering an "always-on" capability without the technical overhead.
Previously, Claude Cowork was tethered to the user’s desktop environment. This meant that any scheduled tasks or ongoing operations would cease if the computer was shut down or the application closed. The new cloud-based architecture ensures that Cowork’s operations, including its ability to execute tasks at predetermined times, will continue seamlessly regardless of the user’s device status. Users can now initiate a task on one device and confidently switch to another, or even disconnect entirely, knowing that their AI assistant is diligently working in the background.
This strategic shift is initially being rolled out to subscribers of Anthropic’s Max plans, which are priced at $100 per month and above. These users gain immediate beta access to the enhanced web and mobile capabilities. Anthropic has indicated plans to extend this functionality to a wider range of subscription tiers in the coming weeks, signaling a commitment to making this powerful tool more broadly available.
A key aspect of this update is the deeper integration of Claude Chat and Claude Cowork. On web and desktop platforms, these two functionalities now share a unified interface. Projects and the artifacts generated by Cowork are readily accessible within the chat environment, fostering a more cohesive user experience. Users can transition between Claude Chat and Claude Cowork through a modal overlay, eliminating the need to navigate to separate sections within the Claude user interface. For mobile users, Claude Cowork will be presented as a distinct section within the dedicated mobile application. Anthropic asserts that this close integration makes delegation as simple as issuing a verbal command, streamlining the process of assigning tasks to the AI.
The company acknowledges that the concept of AI agents remains relatively new for many knowledge workers. When Claude Cowork first launched, there was a notable surge in educational content, including numerous explainer videos on platforms like YouTube, aimed at demystifying its capabilities. This was partly because the familiar back-and-forth chatbot interaction remains the default for most users engaging with AI tools. The introduction of agentic AI, which operates more autonomously to achieve complex goals, represents a paradigm shift. By making Cowork more accessible and seamlessly integrated, Anthropic aims to encourage greater experimentation and adoption among users looking to automate a wider array of workflows and schedule more tasks.
From Claude Code to the Chat Window: Democratizing Agentic Capabilities
At its core, Claude Cowork leverages the sophisticated agentic capabilities initially demonstrated in Claude Code, Anthropic’s specialized coding agent. The innovation with Cowork lies in making these advanced functionalities accessible to a much broader audience, including individuals without technical backgrounds who would typically not engage with command-line interfaces or development environments. Since its launch in January, Cowork has empowered non-developers to delegate tasks to Claude, which can then operate across a user’s files, calendar, email, messaging applications, the web, and other connected tools to achieve the desired outcome.
This evolution represents a significant step towards providing an "always-on" agentic experience that users have been seeking. The previous desktop-only limitation, requiring a computer to remain powered on and the application running, presented a practical barrier. By migrating to the cloud, Anthropic is eliminating this inconvenience and delivering on the promise of continuous AI assistance.
In April, Anthropic officially moved Claude Cowork out of its preview phase and into general availability for its paid plans. This transition also included the integration of enterprise-grade features such as role-based access controls and governance functionalities, addressing requests from business clients seeking to deploy AI agents in a secure and manageable environment. Despite these advancements, the operational constraints of a desktop-bound application persisted until this latest update.
The Work Continues: Cloud-Based Automation and Scheduled Tasks
The most impactful change with this announcement is the shift to cloud-based execution for Cowork’s operations. This means that tasks will continue to run even when the user’s local device is offline. Crucially, scheduled tasks will now reliably execute without user intervention or the need to keep a computer running. This functionality mirrors the capabilities offered by specialized scheduling tools like OpenClaw or hosted services such as Tasklet, but integrated directly into an AI assistant.
This move effectively provides Anthropic users with the "always-on" agent that many desire, without the inherent inconvenience and potential downtime associated with keeping a personal computer operational. For the majority of knowledge workers, this transition bridges a significant gap, offering many of the benefits of advanced automation without the complexity or risks often associated with more technical solutions.
It is important to note a specific caveat regarding existing scheduled tasks. According to an Anthropic spokesperson, these legacy tasks will not be automatically migrated to the cloud. They will continue to run locally and will require the user’s device to remain on. To leverage the new cloud-based scheduling, users will need to create new scheduled tasks or manually move existing ones from the "Scheduled tasks" page. New scheduled tasks created after this update will run automatically in the cloud.
While the cloud offers new avenues for operation, the desktop version of the Claude application retains its relevance, particularly for tasks requiring deep local interaction. Anthropic continues to position the desktop app as the optimal environment for "deep work," primarily due to its ability to access local files and interact with the user’s browser. This raises the question of when cloud-based AI agents might gain ad hoc browser access in the cloud, further expanding their operational scope.
"The Work Around the Work": Unveiling Usage Patterns
Anthropic has also released data offering insights into how Claude Cowork is being utilized, based on an analysis of 1.2 million anonymized user sessions in May. The findings reveal that the vast majority of Cowork usage, over 90 percent, is not related to software development. Instead, the most significant category, accounting for one-third of all sessions, is business operations. This includes tasks such as reconciling spreadsheets, consolidating disparate updates into single reports, and building comprehensive onboarding checklists. Content creation and copywriting represent another substantial area, making up 16.4 percent of usage.
Combined, these "everyday knowledge work" categories account for approximately half of all Cowork interactions. Anthropic defines this type of work as tasks that are "rarely in anyone’s job description, but a large share of everyone’s week." This aligns with the long-standing argument that AI and AI agents have the potential to liberate knowledge workers from repetitive and time-consuming tasks like data compilation and presentation preparation.
Software development, while a prominent use case for AI in general, accounts for a relatively smaller portion of Cowork’s current usage at 8.7 percent. This figure is noteworthy, especially given the proximity of Claude Code within the same desktop application. As Anthropic points out in its report, "While coding is still – understandably – one of the uses of AI that gets the most attention, the use of AI for everyday business work is on the rise, and the kinds of tasks people are finding it most helpful for are coming into focus." This suggests a growing recognition of AI’s utility beyond specialized technical fields.
Closing the Gap: Competition in the Agentic AI Landscape
The move to a cloud-based infrastructure also positions Anthropic more competitively against other major players in the AI agent space. Microsoft’s Copilot, which launched in March, has always operated in the cloud and within the secure confines of a customer’s Microsoft 365 tenant. While Copilot is built upon Anthropic’s Claude engine and agentic framework, its cloud-native deployment from the outset offered a distinct advantage in terms of always-on accessibility.
Similarly, Google’s Gemini Agent Mode and OpenAI’s Operator are actively pursuing the same non-developer automation market. By bringing Claude Cowork into the cloud, Anthropic is effectively leveling the playing field and ensuring its offering is comparable in terms of continuous availability and accessibility.
To commemorate this significant product update, Anthropic is extending doubled Cowork usage limits for all users through August 5. This promotion allows users to more thoroughly test the capabilities of the cloud-based Cowork, including its ability to "keep going without you," a core promise that is now demonstrably achievable. The expanded accessibility and enhanced functionality are expected to drive further adoption and unlock new efficiencies for knowledge workers across various industries.
