Automation

Anthropic's Claude Cowork Now Runs AI Tasks Without Open Laptop

The company's agent can now execute scheduled tasks overnight and work through mobile and web interfaces independently.

Omega Editorial· July 7, 2026· 3 min read

Anthropic announced Tuesday that Claude Cowork, its AI agent for automating digital tasks, no longer requires users to keep their computers running. The agent can now operate independently through smartphone and web browser interfaces, executing scheduled tasks even when desktop sessions are closed.

The update eliminates a significant friction point that previously forced users to leave laptops open overnight to maintain agent sessions. Cowork can now handle requests like compiling data from email threads, Slack channels, and meeting transcripts to generate reference documents and draft communications—all while users are offline.

How the expanded access works

Anthropic previously offered a Dispatch feature that let users send task requests from their phones to desktop-paired agents. However, that approach required the computer to remain awake with the application open. The new implementation removes this constraint entirely, allowing Cowork to function as a truly asynchronous assistant.

Limited versions of Cowork are now accessible through Claude's existing smartphone app and web browser, no longer requiring a desktop connection. Users can submit requests from any device and check results later, rather than maintaining an active session.

Why it matters

This release reflects a broader industry shift toward always-available AI agents that operate more like delegated assistants than interactive tools. For business users, the ability to assign overnight tasks—like analyzing late-arriving messages or preparing morning briefings—could meaningfully change how knowledge work gets scheduled. The move also signals that AI companies see mobile-first, asynchronous interaction as the path to mainstream adoption beyond developer audiences.

Usage patterns and availability

According to a usage report released alongside the announcement, white-collar workers are increasingly integrating Cowork into regular workflows. The two largest use categories are business process and operations tasks (data reports, checklists) and content creation (slide decks, partnership proposals).

The revamped Cowork will roll out in beta to Max plan subscribers, which costs $100 monthly. Features are expected to reach Pro tier members ($20 monthly) later, though availability for free users remains unclear.

Competitive context

The announcement follows similar moves across the AI industry. The trend gained momentum from OpenClaw, a viral homebrew agent that ran continuously for early adopters. OpenAI subsequently hired OpenClaw's creator and launched Codex, its adaptive agent, while Google introduced Spark. OpenAI rolled out Codex Remote in June, enabling smartphone control of desktop agents, and added task management features to its iOS app in July.

Anthropic's approach merges the Claude chatbot interface with Cowork agent capabilities across browser and desktop versions, integrating automation directly into the conversational interface millions already use.

These details were first reported by WIRED.

#anthropic#claude#ai agents#automation#enterprise ai#mobile ai

This is an original analysis by the Omega editorial team. Source reporting: WIRED.

Want systems like this working for your business?

Book a Call

More in Automation

Automation· 3 min read

Rewst's AI Agent Lets MSPs Build Automation Without Deep Coding

New agentic capabilities arriving mid-2025 promise to democratize workflow automation for small business clients through natural language instructions.

Via Automation Watch · Jul 7, 2026
Automation· 3 min read

Siemens Eigen Agent Automates PLC Engineering Tasks in TIA Portal

New AI assistant executes project-aware automation workflows, from ECAD integration through drive configuration, targeting machine builders facing capacity constraints.

Via Automation Watch · Jul 7, 2026
Automation· 3 min read

Summer.fi Vault Exploit Exposes AI Automation Risk in DeFi

A $6 million attack on automated yield vaults reveals how AI-driven rebalancing creates new trust boundaries users rarely examine.

Via Automation Watch · Jul 7, 2026