Anthropic Claude Duet Runs Business Tasks Unprompted
Anthropic has unveiled Claude Duet, a dedicated application designed to autonomously manage enterprise work across everyday communication stacks. Powered by the company's Opus 5.5 model, the tool monitors connected accounts like Gmail, Slack, and calendars to triage tasks without waiting for user commands.
Proactive monitoring without manual prompting
Claude Duet operates as an active background agent rather than a passive chatbot waiting for text input.
Instead of sitting empty until you type an instruction, Claude Duet continuously scans linked workplace channels to flag what matters. The application hooks into services like Slack, Gmail, and scheduling tools, organizing action items based on urgency and advising users on which messages can safely be deferred.
That hands-off posture addresses a frequent complaint in enterprise software adoption: chat-based AI often adds overhead when users must micromanage every query. Recent studies show that checking workplace AI can take longer than writing from scratch, making autonomous background sorting a logical pivot for developers trying to prove direct return on investment.
Under the hood, Duet runs on Anthropic's Opus 5.5 architecture. The model powers the system's contextual understanding, letting it evaluate long email threads or multi-channel Slack discussions to piece together situational context before offering recommendations.
Anthropic Claude Duet preview requirements and details
| Spec | Detail |
|---|---|
| Release status | Research preview (waitlist) |
| Required subscription | Pro or Max tier |
| Underlying model | Opus 5.5 |
| Key integrations | Gmail, Slack, Calendar, productivity apps |
| Target market | Business owners and enterprise professionals |
Privacy boundaries and sensitive data filters
Giving an AI agent read and write access across an entire professional toolchain introduces obvious corporate compliance risks. Anthropic claims it designed Duet around strict data isolation protocols.
The software automatically rejects explicit credentials and payment records, including government identification data and credit card details. When it encounters broader confidential material—such as medical discussions or operational finance logs—Duet pauses its autonomous behavior and demands manual authorization before routing or transmitting that information anywhere else.
Anthropic targeting enterprise automation with Opus 5.5 makes strategic sense, but granting persistent hooks into company-wide Slack and inbox channels will test corporate IT comfort. The hard stop on credit card numbers and identity records is necessary, yet IT managers will likely demand verifiable audit logs before letting Duet scan sensitive internal communications.
Availability and pricing
Claude Duet is currently restricted to a research preview. Anthropic is limiting waitlist registration to existing subscribers on its paid Pro and Max tiers across supported regions, including North America and Europe.
A rollout timeline for free accounts or broader enterprise organization licensing has not been announced.
Frequently asked questions
Who can use Claude Duet right now?
Claude Duet is available through a research preview waitlist open exclusively to Anthropic Pro and Max plan subscribers.
What AI model powers Claude Duet?
Duet runs on Anthropic's Opus 5.5 foundation model, tailored for multi-application context tracking and autonomous enterprise tasks.
How does Claude Duet handle sensitive company data?
It automatically strips payment cards and government IDs. For health or financial data, Duet pauses and asks permission before taking action.
Sources
- Anthropic Introduces Claude Duet: A New Agentic AI App for Professionals - CNET. cnet.com, 2026. https://www.cnet.com/tech/services-and-software/anthropic-claude-duet-ai-agent-app-professional/