Claude Duet Joins The Always-On AI Race As Control Moves To The Switchboard

🔍 Read the full analysis: Claude Duet Joins The Always-On AI Race As Control Moves To The Switchboard on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

Anthropic launched Claude Duet on October 21, introducing a waitlisted research preview of an AI agent that works on a cloud computer and checks connected apps for tasks between user prompts. The service is limited to U.S. Pro and Max subscribers, and its arrival adds Anthropic to a developing race among technology companies selling assistants that can keep working in the background.

Duet is a separate app available through iOS, Android and claude.ai, according to Anthropic. Users can give it tasks by text or voice in an ongoing conversation. The company says it can prepare a morning brief, schedule a meeting, or build items such as a dashboard, sales forecast or small web app. Work runs on a virtual machine with its own browser, so the user’s phone or computer does not need to remain on.

Anthropic says Duet carries over the memory, skills and connected apps associated with a user’s Claude account and can draw on more than 6,000 connectors and plugins in the Claude Marketplace. Its examples include finding unpaid invoices in QuickBooks or Stripe, payroll information in Gusto, agreements in DocuSign and storefront tasks in Shopify. For purchases, Anthropic says it uses Stripe Link with a one-time card restricted to a specified merchant and amount.

Anthropic’s October 23 safety post says Duet must show users the action, destination and content before sending a potentially sensitive message, making a payment or deleting something. It waits for approval, while a separate classifier checks whether an action matches the user’s intent. The company says trading securities, moving money between accounts and accessing another person’s account are blocked even with user approval. These are Anthropic’s descriptions; the source material reports no independent tests of the preview.

Control Over the Agent’s Switchboard

Duet’s launch is part of a broader shift from chatbots that respond on demand to agents that retain context and act in the background. For users and businesses, that could mean less manual coordination across email, billing, scheduling and other connected services. It also means granting an AI system access to more accounts and deciding which actions it may take without direct supervision.

The source frames the emerging competitive question as who controls the interface, connected data and routing of work, rather than which company supplies every model. It says Grok Bot is expected to route tasks to third-party models, including Claude Opus 5.5, and that its model choice cannot be restricted through a customer-facing picker. If that account is accurate, an assistant provider may determine which model handles a task even when another company’s model performs the underlying work. The source does not establish how broadly this arrangement will operate or how customers will be informed for each task.

For organizations considering adoption, the practical stakes are permissions, oversight and cost. Duet’s background checks count against the same usage limits as other Claude activity, according to Anthropic. A pilot therefore needs to measure not just time saved but also review effort, usage and the consequences of mistakes. The source recommends small pilots with limited permissions and a clear exit plan; that is guidance, not a reported company policy.

A Crowded Field of Background Agents

The source places Duet among several launches in a short period. It says SpaceXAI opened the Grok Bot beta on August 11, Meta launched Muse on September 8, OpenAI introduced Dots at DevDay on September 29, and Google released Gemini Agent on October 8, citing TechTarget for Google’s date. Anthropic followed on October 21. Each product’s capabilities and availability may differ, so the launch sequence alone does not establish a direct feature-by-feature comparison.

Anthropic presents Duet as a tool for professionals and small-business owners managing work across connected services. The source says the preview runs on Claude Opus 5.5 and is restricted to U.S. Pro and Max users through a waitlist. It does not provide independently verified performance comparisons, adoption figures or a complete account of rival products’ current terms and capabilities.

“for professionals and small business owners with more on their plates than they can track.”

— Anthropic, describing the intended audience for Claude Duet

Testing, Access and Data Retention

Duet remains a waitlisted research preview, and the source says no independent testing is available. Its reported safety controls and capabilities are based on Anthropic’s own statements, so their performance in real-world use has not been established here. The source also does not specify how long the waitlist will remain in place or when access may expand beyond eligible U.S. subscribers.

Anthropic says Duet checks connected accounts every few hours rather than continuously, to limit usage. That could leave gaps between checks, while background activity and larger tasks use the user’s Claude plan limits. The company also says disconnecting an app prevents new access but does not erase information Duet has already learned; users must reset its memory to clear that information. The source does not detail the full retention period or all data-handling terms.

Other limits include weaker performance on personal errands such as booking tickets and no current way for users to view Duet’s browser, according to the source. It remains unclear how often approval prompts will appear in routine use, how reliably the separate classifier will catch actions that do not match intent, and whether the product’s safeguards will change during the preview.

How the Preview Will Be Evaluated

The immediate next step is access through Anthropic’s waitlist for eligible U.S. Pro and Max subscribers. The source provides no public timetable for a broader release or a full launch. Users who gain access can test routine, low-stakes tasks while keeping approvals active for actions involving messages, payments or deletion, consistent with the controls described by Anthropic.

Businesses evaluating Duet can start with a limited set of connected services and written rules for what the agent may read or prepare. They can track errors, approval requests, time spent reviewing work and usage against plan limits before expanding access. Those measures would help establish whether the system is useful in a particular workflow; the source reports no completed enterprise trials or measured productivity results.

Key Questions

What is Claude Duet?

Claude Duet is Anthropic’s cloud-based AI agent, offered as a separate app and through claude.ai. Anthropic says it can review connected apps, send a morning brief and carry out assigned work in the background.

Who can use the research preview?

The source says access is limited to U.S. Pro and Max subscribers and is provided through a waitlist. It gives no date for wider availability.

Can Duet send messages or make payments on its own?

Anthropic says Duet must request approval before sending potentially sensitive messages or making payments. It says some actions, including trading securities and moving money between accounts, are blocked even if a user approves them.

Does Duet monitor connected accounts continuously?

No. Anthropic says Duet checks connected accounts every few hours, not nonstop, partly to manage usage. That means it may not detect a new item between checks.

Has Duet’s performance been independently verified?

The source reports no independent testing of the preview. Descriptions of its features, safeguards and limitations are attributed to Anthropic.

Source: ThorstenMeyerAI.com