TITLE: Claude Cowork Goes Mobile: What Agents Actually Do at Work DATE: 2026-07-12 COMPANY: Anthropic TOPIC: Agent Systems SUMMARY: Anthropic expanded Claude Cowork to mobile and web on July 7, 2026, letting AI agents run tasks in the background even when a user's laptop is closed. In doing so, Anthropic released usage data that challenges the coding-first narrative around AI agents: more than 90% of Cowork sessions involve non-coding tasks, with business process operations (33.4%) and content creation (16.4%) leading the way. WHAT CHANGED: Anthropic released Claude Cowork as a desktop application in January 2026, positioning it primarily as an agentic coding tool. On July 7, the company began rolling out web and mobile access, starting with users on the Max plan and expanding to additional tiers over the following weeks. The core capability that makes mobile access meaningful is background persistence. Previously, a Cowork session required the user's device to remain active. With the new architecture, tasks can be started on a desktop, continue running in the background through Anthropic's infrastructure, and be monitored or retrieved on a mobile device. The agent keeps working even when the laptop is closed. Alongside the launch, Anthropic released an analysis of how users are actually using the product. The data showed that more than 90% of all Cowork sessions involve tasks that are not software development. The largest category, at 33.4%, was business process operations: pulling updates from scattered sources into a single report, building onboarding checklists, and reconciling spreadsheets. Content creation and copywriting, including drafts, slide decks, social posts, and proposals, made up 16.4%. Anthropic has positioned the mobile and web expansion as a signal that Cowork is a general-purpose business productivity tool, not a developer product. The platform now unifies Chat and Cowork in a single interface on web and desktop, with projects and artefacts accessible across both. --- WHY IT MATTERS: AI agents have moved past the developer desk. The 90% non-coding usage figure matters because it directly challenges the common rollout pattern where AI tools land in engineering first and expand outward later. Anthropic's data suggests the demand and volume are already sitting in operations, communications, and content teams. Background execution changes the task economics. Any agent-driven task that used to require a babysitter now runs unattended. This is significant for longer workflows: overnight reports, multi-source research briefs, or document preparation that runs during a meeting and delivers output when the team is ready for it. Mobile access removes a friction point for non-technical users. The people doing business process operations and content work are not sitting at a development workstation. Bringing Cowork to mobile means the tool fits into the way those users actually work, rather than asking them to adapt to a desktop-first tool. The pricing gate has implications for teams. Background agent access on mobile is currently restricted to Max subscribers at $100/month. For teams wanting to deploy this capability across multiple staff members, the per-seat cost is a real factor to plan for. Enterprise plans with custom pricing offer a path for larger deployments. Anthropic is repositioning Cowork as an operating layer. By revealing usage data and framing the launch around business process categories rather than coding features, Anthropic is signalling the product direction. The addressable market for Cowork is not just developers; it is every team in a company that produces documents, reports, and communications at scale. The workflow pattern has implications for how work is structured. When agents run in the background and deliver finished outputs, the human role shifts from doing the task to reviewing and approving it. This is a structural change in how work gets organised, and operators who design their teams around it will have a meaningful productivity advantage over those who treat AI as an add-on. --- DAVID & GOLIATH ANALYSIS: The usage data Anthropic released is the most strategically important part of this announcement. For the past two years, the dominant narrative around AI agents has been developer-centric: code generation, debugging, automated testing, software pipelines. That framing made it easy for operators without large engineering teams to defer AI agent adoption. It also made the ROI case feel contingent on technical headcount rather than business volume. The Cowork numbers flip that logic. The highest-volume use case is not writing code; it is reconciling spreadsheets, building onboarding checklists, and assembling status reports. These are tasks that exist in almost every team inside a 10-200 person company, often done by people who have no interest in or access to developer tools. The implication is that the best place to start an AI agent rollout is not your tech team. It is your operations manager, your marketing coordinator, and your account manager. Background persistence is the capability that makes the economics work at that layer. A report that takes four hours of fragmented attention, pulling from emails, shared folders, and spreadsheets, can now be handed to an agent at the start of the day and retrieved as a completed draft before lunch. The agent does not need oversight while it runs, and the user does not need to be at their desk. That changes what is worth automating and how organisations should think about where AI fits into the daily rhythm of their teams. --- RELEVANT SYSTEMS: Employee Amplification Systems, AI Growth Engine SOURCE URL: https://davidandgoliath.ai/daily-ai-briefing/anthropic-claude-cowork-mobile-web-background-agents FEED URL: https://davidandgoliath.ai/daily-ai-briefing/feed --- Published by David & Goliath | https://davidandgoliath.ai Daily AI Briefing: one AI development per day, decoded for business operators. This is a structured companion file optimised for LLM retrieval and citation.