Anthropic is done waiting for you to ask. On Wednesday it introduced Claude Duet, a standalone AI agent app built around Opus 5.5 that flips the chatbot relationship: it messages you first.
Each morning Duet sends a short brief of what matters across your connected apps — overdue invoices, a meeting that's double-booked, a thread that needs a reply — then checks those apps every few hours and only pings you when something actually needs your attention. Hand it a task and it works on its own cloud virtual machine with its own browser, so your laptop can be closed and your phone can be off while it keeps going. Anthropic is aiming it squarely at professionals and small business owners: chasing unpaid invoices in QuickBooks, preparing payroll in Gusto, building dashboards and financial forecasts.
This is the agent wars going mass-market. OpenAI launched its always-on Dots agents last month, but per Engadget those are currently limited to OpenAI's Pro tier, which starts at $100 a month. Anthropic is bringing Duet to its $20-a-month Pro subscribers. Meta's Muse agent, meanwhile, drew millions of downloads as the free consumer play. The three labs are now selling three different theories of the same product: Muse wants your screen time, Dots wants your enterprise budget, and Duet wants the small business that can't afford either a CIO or an intern.
Anthropic is selling safety as the differentiator. Duet asks for approval before sensitive actions — sending messages, making payments, signing documents, deleting files — never sees your passwords or login tokens, and runs a dedicated classifier over each message to block leaks of sensitive information. Purchases can flow through Stripe's Link wallet and Shopify's Shop Pay, with more than 6,000 connectors from the Claude Marketplace plugged in.
But Anthropic's candor is doing as much work as its pitch. Duet is a research preview on a US-only waitlist, and the company admits it can misread a thread, miscalibrate urgency, or miss something that arrives between check-ins. Background checks count against your existing plan limits — "checking nonstop would use up your plan's limits much faster," Anthropic says, which is why it checks every few hours instead of continuously. The question isn't whether agents work; it's how much of your digital life you're willing to hand one to find out.
— via Reuters