Claude Duet: The AI Agent That Works While You Sleep

The Out-of-Hours Problem Is Costing You Business

Every estate agency faces the same uncomfortable truth: enquiries do not stop when your team goes home. A prospective buyer submits a viewing request at 9pm. A tenant flags an urgent maintenance issue on Saturday morning. A vendor chases for an update on a Friday afternoon. By Monday, the moment has passed.

The cost is not just a lost lead. It is the compounding effect of consistently slow responses on your reputation, your pipeline and your negotiating position. The agencies winning instructions right now are not necessarily the biggest — they are the fastest and most attentive.

For most firms, the gap between enquiry and response is filled by manual effort: someone checking emails out of hours, someone drafting the same follow-up for the hundredth time, someone deciding what is urgent and what can wait. That is an expensive, unsustainable way to compete.

• Enquiries received outside office hours frequently go unanswered until the next working day
• Staff time is consumed by repetitive monitoring and drafting tasks rather than client-facing work
• Slow response times directly affect conversion rates at every stage of the pipeline
• Compliance-sensitive communications require careful handling, slowing down even routine correspondence
• Inconsistent follow-up quality creates risk and undermines client confidence

Why the Old Fixes Have Not Worked

Agencies have tried to solve this with chatbots, auto-responders and out-of-hours call handling. These tools manage expectations but they do not advance the work. An auto-reply tells a client you received their message. It does not draft the follow-up, check the diary for availability or flag the enquiry as high priority in the CRM.

The deeper problem is that most AI tools available to agencies until now have been reactive. They wait to be asked. Someone has to open the tool, type the prompt and review the output. That still requires human attention — which is exactly the resource in short supply outside working hours.

• Reactive AI tools require a human to initiate every task, limiting their value out of hours
• Chatbots create the appearance of responsiveness without advancing the underlying work
• Disconnected systems mean monitoring, drafting and sending happen in separate manual steps
• Fear of compliance errors causes teams to over-supervise AI output, negating time savings

What Forward-Thinking Agencies Are Already Doing

The agencies setting the pace are not waiting for a perfect solution. They are mapping their workflows now — identifying which tasks are genuinely about monitoring and drafting versus which require real professional judgement. That distinction matters enormously as proactive AI agents move from research previews into production tools.

Anthropic has previewed Claude Duet, a proactive work agent available to US Pro and Max subscribers, that monitors connected applications, prepares work and completes multi-step tasks in the background — pausing to request human approval before taking sensitive actions such as sending messages, making payments or deleting files (MacRumors, 2026). Forward-thinking businesses are paying close attention to how the approval-gate design keeps humans in control of consequential decisions while removing the administrative burden from routine ones.

• Auditing current workflows to separate monitoring and drafting tasks from genuine judgement calls
• Identifying which communications are compliance-sensitive and therefore require an approval gate
• Testing proactive AI tools in lower-risk workflows before applying them to client-facing correspondence
• Building internal familiarity with agent-based AI so staff can supervise output confidently

How Proactive AI Agents Change the Operating Model

The shift from reactive chatbot to proactive agent is not a minor upgrade — it changes which tasks require human attention and when that attention is needed. Rather than doing the work, your team reviews and approves it.

Background Monitoring and Triage

A proactive agent connected to your inbox, calendar and CRM can check incoming enquiries every few hours, without anyone having to open a dashboard. It surfaces what matters and deprioritises what does not.

• Checks connected applications every few hours and sends an alert when something needs attention (MacRumors, 2026)
• Flags urgent enquiries for human attention, whenever they arrive
• Categorises incoming messages by type, sender and required action
• Prepares a prioritised task list ready for the first person online each morning

Drafting and Preparation

The agent does not just identify what needs a response — it prepares one, ready for approval. This is the part of the job an agent can take off your team's plate.

• Drafts follow-up messages based on enquiry content and previous correspondence
• Prepares calendar invitations and availability options for viewing requests
• Pulls relevant property details or tenancy information into draft responses automatically
• Queues all drafts for a single human review session rather than scattered manual effort

Approval Gates for Sensitive Actions

The detail that makes this model viable for compliance-conscious agencies is the approval gate. No message is sent, no file deleted, no payment initiated without a human sign-off (MacRumors, 2026).

• Holds all outbound communications until explicitly approved by a team member
• Flags which actions are sensitive and explains why approval is required
• Keeps compliance-critical communications — offers, tenancy notices, financial referrals — under direct human control
• Creates an auditable record of what was prepared, reviewed and approved

A Practical Roadmap for Agencies Ready to Act

Adopting proactive AI does not require a wholesale technology overhaul. The agencies that move quickest will be those that start with clarity about their own processes before selecting any tool. Here is a sequenced path forward.

• Map your current workflows and separate monitoring-and-drafting tasks from those requiring professional judgement — this distinction determines where agents will add immediate value
• Identify which outbound communications are compliance-sensitive and therefore must remain behind an approval gate under any AI-assisted model
• Select one lower-risk workflow — such as viewing request acknowledgements — to pilot a proactive agent approach before extending to client-facing correspondence
• Brief your team on the approval-gate model so they understand their role shifts from doing the work to reviewing and releasing it
• Establish a review cadence: how often will a team member check the approval queue, and who is responsible out of hours for genuinely urgent items

The Business Case Is Already Clear

An agent that monitors your inbox, prepares follow-up drafts and flags urgent enquiries — without being prompted, without waiting for office hours — changes the economics of running an estate agency. The administrative groundwork gets done whether or not anyone is at a desk. Compliance remains in human hands. And your team arrives each morning with work already prepared rather than a backlog to wade through. The firms that understand this now, and act on it, are the ones that will set the terms for everyone else.

🚀 Find Where Your Enquiries Leak

Start with a free Revenue Leak Audit. We score six points where enquiries can leak, from your Google reviews to how enquiries get followed up, and send you a scorecard within 48 hours: aiviasystems.ai/free-revenue-leak-audit

To talk it through, email [email protected].

References

• MacRumors. (2026). Anthropic Claude Duet. MacRumors. https://www.macrumors.com/2026/10/21/anthropic-claude-duet/