185 open opportunities · 21 added this week · 2 closing within 7 days See what's open
Back to all news

Anthropic Launches Claude Duet, an AI Agent That Works in the Background and Manages Tasks

Anthropic has introduced Claude Duet, a proactive AI agent designed to help professionals and small-business owners manage work across their connected applications. The tool can prepare daily briefings, monitor email and calendars, draft communications, create dashboards and complete tasks in the background. However, access remains limited to a research preview for Claude Pro and Max subscribers in the United States who join the waitlist.

Anthropic Launches Claude Duet, an AI Agent That Works in the Background and Manages Tasks

Anthropic Introduces Claude Duet, an AI Agent Designed to Work While You Are Away

Artificial intelligence company Anthropic has introduced Claude Duet, a new AI agent designed to help professionals manage their workload without having to instruct a chatbot at every step.

Announced on October 21, 2026, Claude Duet is built to monitor connected applications, identify information that requires attention and work on tasks in the background. Instead of waiting for users to open a chat and ask questions, the agent can send a morning briefing, prepare documents, organise information and notify users when a task is complete or a decision is needed.

Anthropic is targeting professionals and small-business owners who manage multiple responsibilities and want to reduce the time spent on routine administrative work.

The launch marks another development in the growing market for AI agents: systems that do more than generate text and can take action across connected software.

However, the product is still in its early testing stage. Access is currently available by waitlist to Claude Pro and Max subscribers in the United States.

What Is Claude Duet?

Claude Duet is a proactive AI agent that works through messages and connected applications.

A traditional AI chatbot generally responds when a user asks a question or gives an instruction. Duet is designed to take a more active role by checking connected accounts periodically, identifying important developments and preparing work before the user requests it.

For example, the agent may review a calendar, notice an upcoming client meeting and prepare a briefing using relevant information from connected applications. It may also identify a scheduling conflict and suggest a change.

For a business owner, it could help review unpaid invoices, prepare payment reminders, organise financial information or create a dashboard showing business performance.

Anthropic says Duet is powered by Claude Opus 5.5 and builds on the memory, skills and connected applications available through a user's Claude account.

The objective is to help people spend less time coordinating routine work and more time making decisions, solving problems and handling responsibilities that require human judgement.

What Can Claude Duet Do?

Anthropic has described several capabilities for the new agent.

1. Prepare a daily briefing

Claude Duet can prepare a morning summary of important developments across connected applications.

It may highlight significant email conversations, upcoming meetings, pending decisions and tasks that require attention.

Instead of manually opening several applications to understand what is happening, users can receive a consolidated overview of outstanding work.

2. Monitor connected applications

The agent can connect to services through Claude's connectors and plugins.

Anthropic says the system supports more than 6,000 connectors and plugins. These can include tools for email, calendars, customer relationship management, accounting, project management and document preparation.

Duet checks connected accounts every few hours during the research preview, rather than continuously monitoring them. This helps manage resource usage while allowing it to identify developments that may require attention.

3. Draft emails and follow-up messages

Duet can examine relevant conversations and prepare suggested replies.

For businesses, it can also use available customer or deal information to prepare follow-up emails. This could help reduce the time spent writing routine communications and gathering background information.

Importantly, the system is designed to seek approval before sending messages on a user's behalf.

4. Build dashboards and business reports

The agent can work on more complex tasks, including preparing dashboards, creating financial or sales forecasts and organising information from multiple tools.

A small-business owner might use it to combine sales, customer support and accounting information into a single overview.

These capabilities could be useful for people who do not have the time or technical expertise to build every report manually. However, users still need to check the data, assumptions and conclusions before relying on the output for important decisions.

5. Support everyday business operations

Anthropic's examples include helping users prepare payroll information for review, identify unpaid invoices, draft job advertisements, organise agreements and prepare documents for signature.

The product can also support work involving platforms such as Shopify, HubSpot, Stripe, QuickBooks, Google Drive, Gmail and Slack, depending on the available integrations.

The aim is to connect these tasks within one working environment instead of requiring users to move repeatedly between different applications.

6. Continue working when the user is away

One of Duet's main differences from a conventional chatbot is that it has access to a virtual machine with its own browser in the cloud.

This allows it to continue working after a task has been assigned, even when the user's laptop or phone is not active.

When the work is finished or a decision is required, the agent can send a notification.

Anthropic describes this as an assistant that brings important information to the user, rather than requiring the user to initiate every interaction.

How Does Claude Duet Differ From a Regular AI Chatbot?

The difference is primarily how the system approaches work.

A conventional chatbot is usually used in a question-and-answer interaction. The user gives an instruction, receives a response and decides what to do next.

An agent such as Claude Duet is designed to operate across a longer sequence of tasks. It can gather information from connected applications, prepare a result, continue working in the background and return when the task is complete or approval is needed.

For example, a user could ask a conventional chatbot to explain how to prepare a sales report. With an agent, the intended workflow is more direct: the system can access authorised data sources, assemble a report and present it for review.

This does not mean that the agent can reliably handle every task without supervision. It means the product is designed to carry out more of the steps between an instruction and a completed piece of work.

The distinction is important because AI companies are increasingly competing to build systems that can perform useful actions, rather than simply provide information.

Safety, Privacy and Human Approval

Giving an AI agent access to email, calendars, accounting platforms and business documents raises important security questions.

A system that can read business information and prepare actions on a user's behalf must be designed to prevent accidental disclosures, unauthorised transactions and unintended changes.

Anthropic says Claude Duet includes several safeguards.

Approval before consequential actions: The agent is designed to ask for permission before sending messages, making payments, signing documents or deleting information.

Protection for sensitive information: Anthropic says additional safeguards are designed to prevent the sharing of highly sensitive information, including credit-card numbers and government identification. Other sensitive information, such as personal financial or health data, may be shared only when the user has explicitly instructed Duet to do so.

Credential protection: The company says Duet's design prevents the agent from directly seeing users' passwords and sign-in tokens.

User visibility and control: Users can review what the agent is doing and disconnect applications from which it should no longer retrieve information.

These protections are important, but they do not eliminate every risk. Anthropic's documentation acknowledges that the product is still under development, and its safety and security measures are continuing to evolve.

Users should carefully consider which accounts they connect and avoid granting unnecessary access to sensitive information.

Claude Duet Is Still in a Limited Research Preview

Despite the product's broad capabilities, most users cannot simply download it and begin using it.

Anthropic initially made Claude Duet available through a waitlist to Claude Pro and Max subscribers in the United States. The company has not announced general availability for every user or confirmed access for all countries and subscription plans.

The product is available as a standalone application for iOS and Android and through the web at Claude.ai.

During the research preview, Duet shares usage limits with the rest of the user's Claude plan. Background checks and complex assignments also consume usage, meaning larger tasks can use more of the available allowance than a typical chat.

Anthropic has also acknowledged performance limitations. Duet may misunderstand a conversation, misjudge which tasks are urgent or fail to identify something that happens between its periodic checks. Its current design is focused more on professional work than personal errands such as booking travel.

These limitations mean that users should treat the product as an early-stage assistant rather than a fully reliable replacement for human decision-making.

Anthropic Enters the Competition for Personal AI Agents

Claude Duet enters an increasingly competitive market for AI agents that can handle work across multiple applications.

OpenAI and Meta are also exploring more proactive AI experiences through their own products and initiatives. These systems aim to reduce the number of steps users must perform manually by allowing AI to organise information, complete tasks and take action within approved limits.

Anthropic's approach places emphasis on professional workflows and small-business operations rather than presenting the product primarily as a digital companion.

Its launch also reflects a wider shift in AI development: moving from tools that generate content towards systems that can carry out multistep work.

The competitive challenge will be demonstrating that these agents are not only impressive in demonstrations but also dependable, secure and useful in real working environments.

What Could Claude Duet Mean for Businesses and Workers?

If AI agents become reliable and widely available, they could change how professionals manage everyday work.

Small businesses may be able to automate parts of reporting, customer follow-up, scheduling and administrative coordination without building custom software for each process.

Product managers, consultants, sales teams and operations professionals could use agents to gather information, prepare meetings, track tasks and organise business data.

However, the introduction of such tools may also increase the importance of checking outputs, protecting confidential information and deciding when an AI system should act independently.

The value of an AI agent depends not only on what it can generate, but also on whether it can complete a task correctly, respect permissions and handle unexpected situations.

For workers, learning how to delegate tasks to AI and verify its results may become an increasingly useful skill. For companies, the challenge will be deciding which workflows can safely be automated and which still require direct human oversight.

Claude Duet is too new to establish how much time or money businesses will save in practice. Those benefits will need to be demonstrated through real-world use.

Conclusion

Anthropic's Claude Duet represents a move towards AI systems that can monitor connected applications, prepare work and carry out tasks in the background.

Its ability to prepare daily briefings, organise business information, draft communications and build reports could make it useful for professionals and small-business owners managing several responsibilities at once.

Yet the product remains a limited research preview, with restricted access and acknowledged performance limitations. Its long-term value will depend on how reliably it performs, how well its safeguards work and whether users are comfortable giving it access to their applications.

For now, Claude Duet offers an early look at a future in which people may spend less time coordinating routine tasks and more time reviewing results and making decisions.

The key question is no longer just how well AI can answer a prompt, but how reliably it can complete useful work while keeping the user informed and in control.

Official Sources

Anthropic's Claude Duet announcement: https://claude.com/resources/articles/claude-duet

Claude Duet product page and waitlist: https://claude.com/product/claude-duet

Gizmodo's report: https://gizmodo.com/anthropics-new-agent-has-no-cute-mascot-and-you-probably-cant-try-it-yet-2000829184

Never miss a new opportunity

Get instant alerts, join our WhatsApp community, or take a quick break with a game.