Your next hire may not wait for instructions.
Anthropic opened a waitlist preview of Claude Duet on 21 October 2026: an app where the AI messages you first with what needs attention, then keeps working after you close the laptop.
The details matter. Duet sends a morning brief drawn from your connected apps, checks those accounts every few hours, and runs on its own virtual machine with its own browser, so tasks continue while your phone is off. Before it sends a message, makes a payment, signs a document or deletes anything, it shows you the action and waits for approval. Trading securities and moving money between accounts are blocked even with approval.
The preview starts with US Pro and Max subscribers. Australian access has no announced date yet.
For business owners, the shift is not the chatbot. It is the direction of initiative: AI that watches the work, decides what deserves your attention, and asks permission only at the point of action.
The governance question comes with it. Which systems may an agent read? Which actions stay behind an approval? Who audits what it did overnight?
Proactivity gets the headline. The approval boundary is the product.
What would you let an AI start on your behalf before 7 AM?
Think mathematically. Engineer intelligently. Lead strategically. Build responsibly. Scale globally. Create real business value through artificial intelligence.
