Real Estate AI Receptionist: Cusani 10 Case Study
How Milan-based Cusani 10 Luxury Real Estate handled 843 inbound calls and secured 265 agent callbacks using an AI phone assistant.
In a real estate agency, calls arrive while agents are out on property viewings, in negotiations, or meeting with notaries. Inbound callers inquiring about a listing expect immediate answers on availability, asking price, and viewing times. Cusani 10 Luxury Real Estate, a brokerage in central Milan, assigned first-line phone handling to Sara, an AI phone assistant built on DeepAgent. Over six months, the system handled 843 inbound calls, logged 265 agent callbacks, and saved nearly 23 hours of phone time. ## The client: Cusani 10 Luxury Real Estate, Milan ### About the firm Cusani 10 Luxury Real Estate operates from Via Cusani near Milan's Sforza Castle, handling residential sales and acquisitions alongside property valuations and marketing campaigns. The founder has brokered real estate since 2004. ### The operational challenge Published listings generate regular phone inquiries. Callers ask about availability, confirm pricing, verify floor levels, request viewings, or ask for specific brokers. These calls arrive throughout the day, often while agents are away from the office. Missing these calls delays buyer responses. ## The objective: answering property inquiries during agent appointments The objective was to answer calls immediately, identify caller intent, and route qualified contact details to the sales team. The agency set three requirements: - **Accurate listing data.** Callers cite specific addresses or listing codes. The assistant must pull data directly from verified property records rather than giving generic answers. - **Structured lead capture.** When callers request an agent, the assistant must record caller names, phone numbers, and availability windows. - **Uninterrupted agent workflow.** Routine questions must resolve on the call without interrupting client meetings and property viewings. ## How the AI phone assistant operates Sara functions as an inbound voice agent on the main office line, connected directly to the agency's property database. **1. Greets and routes.** The assistant identifies whether the call concerns a property inquiry, a sale, a rental, an appointment, a specific broker, or general office details. **2. Retrieves the listing.** When the caller provides an address or property reference code, the assistant queries the management database and uses verified listing data. **3. Answers questions.** The assistant answers questions on pricing, floor level, elevator access, availability, and office hours directly from the listing file. **4. Hands off qualified leads.** For viewings or detailed requests, Sara records the caller's name, phone number, and preferred callback time. Messages and cancellations route directly to the team inbox. ## Key performance metrics: 843 calls managed | Metric | Result | |---|---| | Calls handled | 843 | | Completed calls | **843 (100%)** | | Transcribed calls | 838 (99.4%) | | Total conversation time | **nearly 23 hours** (22 h 58 min) | | Average call duration | 1 minute 38 seconds | | Calls requesting or scheduling a callback | **265 (31.4%)** | | Calls regarding viewings and appointments | 130 (15.4%) | The system handled 760 calls from June through September (90% of total volume), averaging roughly 190 calls per month. ## Call breakdown and demand patterns Two out of three calls (558 calls, or 66.2%) focused on a property listing or property type. Another 41.2% of calls requested general business details such as office location, business hours, or direct contact lines. Across the dataset, 265 calls concluded with an agent callback request, and 130 calls addressed viewing arrangements. For example, when a caller asks whether a two-bedroom apartment is available for viewings, Sara retrieves the listing, confirms bathroom count, floor level, elevator access, and price, then records the caller's contact details and preferred viewing times. Additionally, 411 calls (48.8%) reached full resolution during the initial call, answering listing queries or passing qualified notes to a consultant. The remaining calls answered general office questions, covered brief routine inquiries, or ended when callers chose not to leave contact details. ## Operational adjustments Implementation indicated several practical refinements: - **Separate KPIs by outcome.** Tracking separate metrics for information provided, viewing requests, callbacks booked, messages taken, and disconnected calls provides clearer operational reporting. - **Concise data delivery.** Initial voice responses after database lookups must remain brief to avoid conversational pauses. - **Clear pricing terms.** When property records list annual rental figures, the assistant should state both the annual figure and the monthly equivalent. - **Downstream pipeline tracking.** Connecting logged callbacks to viewings, offers, and signed mandates tracks commercial conversion. ## Summary An AI receptionist for real estate keeps phone lines covered throughout the day, resolves basic listing questions in under two minutes, and delivers organized contact details for follow-up. For Cusani 10, the system answered 843 calls, organized 265 callbacks, and saved almost 23 hours of phone time for property agents. Talk to sales → · Create your agent →