AI Calling for Real Estate in India: Enquiry to Site Visit Playbook

AI calling for real estate is simple to describe and easy to get wrong. An AI voice agent calls every property enquiry within minutes, qualifies it in the buyer's language, books a site visit and hands a warm lead to your sales team. This playbook covers how we run it for Indian developers — including the mistakes we made and fixed along the way.

What is AI calling for real estate, and where does it fit?

AI calling for real estate means a voice agent, built on your project facts and scripts, that makes and answers calls for you. It calls fresh enquiries, asks qualification questions, answers from approved facts, books site visits and call-backs, and records the outcome. Your sales team handles negotiation and closing.

The case for real estate AI calling in India is about attention, not replacing people. Buyers have choices: ANAROCK data reported by Business Today put July–September 2026 launches in the top seven cities at about 1,14,320 units against sales of about 1,00,220. More supply than sales means every enquiry is also looking at other projects, and the developer who calls first and clearly gets the first site visit.

What the agent does well

• Calls every new lead within minutes, including at 9 PM on a Sunday when the ad is running and your team is not.
• Asks the same qualification questions every time and writes the answers down.
• Speaks Kannada, Hindi, English, Hinglish, Marathi, Tamil, Telugu or Bengali, switching to the caller's language.
• Books visits and call-backs at the time the buyer asks for, and sends a WhatsApp confirmation through your own WhatsApp provider.
• Answers inbound calls after hours without making anything up.

Where AI calling is the wrong tool

It is the wrong tool for price negotiation, for a buyer who has already paid a booking amount and has a complaint, for a long-standing investor client who expects your senior person, and for any question your locked facts do not cover. In each of those, the agent's job is to take details and book a call with a human. A good AI calling setup for real estate is designed around that handover, not around avoiding it.

Where do real estate leads come from, and how fast should you call them?

Most Indian developers get enquiries from Facebook and Instagram lead ads, property portals, brochures and QR codes, events and walk-ins. Each source should feed one list the agent reads automatically. Call new enquiries within minutes: the buyer filled your form moments ago, and every hour of delay lets other projects reach them first.

At Koravi Developers, a plotted-development company in North Karnataka we work with, leads come from Facebook/Instagram ads, brochures, events and walk-ins. That mix is typical, and each source behaves differently, which is why using AI for real estate leads starts with connecting every source to one list.

Speed to lead: why minutes matter

We check connected Google Sheets every 2 minutes and pick up Facebook/Instagram lead ad entries as they arrive, so the first call usually lands while the buyer still has your ad in mind. Research on sales leads has made the same point for years: a Harvard Business Review study of online leads found most companies were "not responding nearly fast enough". In property, slow response means the buyer meets a competitor's sales person first.

Our guide on calling Facebook leads automatically from Google Sheets covers the setup step by step.

What should the first call sound like?

The first call should confirm who the buyer is and what they enquired about, ask three or four qualification questions, answer one or two questions from approved facts, and end with a booked site visit or call-back. Keep each agent turn to one or two sentences, and use the buyer's name with "ji" or "avare".

Here is the flow we use for plotted and apartment projects:

1. Open in the right language. The opening line plays in the project's default language: Kannada for a Hubballi project, Hindi for a Delhi NCR one. We once had an agent go silent because its opening language was set wrong; check this before launch.
2. Say who you are and why you are calling. Name the project and the enquiry: "You asked about plots at [project] on Instagram this morning."
3. Confirm it is a good time. If not, book a call-back at the time they name.
4. Ask the qualification questions (next section), one at a time.
5. Answer from locked facts only. One number per turn (price, then size, then location) never three figures in one breath.
6. Offer a site visit with two specific options. "Would Saturday morning or Monday afternoon suit you?"
7. Confirm the details back. Date, time, pickup point if needed, number of people.
8. Close with a proper goodbye. Our platform holds the hang-up until the goodbye has finished playing; a call that cuts off mid-sentence feels like a scam.

What our callers taught us about tone

A sales manager who tests our calls gave us feedback that changed how every real estate agent sounds:

• Long monologues feel rushed, even when the voice speed is fine. One or two sentences per turn, then stop and let the buyer speak.
• Repetition comes from long turns. When a buyer says "hello?" mid-sentence, a naive agent restarts its whole line. Shorter turns cure most of it.
• Never assume gender. Calling a woman "Sir" repeatedly is a real failure on Indian calls. Use the name with "ji" or "avare".
• Switch language on a full sentence only. An agent that says the same thing in Kannada and then in English back to back sounds broken. Switching on a complete sentence, not on a stray "hello" or "haan", fixes it.

At Koravi, agents call enquiries in Kannada, Hindi and English, because buyers in North Karnataka move between all three. Our article on multilingual AI calling in Hindi, Kannada, Tamil and Telugu goes deeper into language handling.

Which qualification questions matter for property buyers?

Ask only what changes your next step: what they want, where, at what budget, by when, whether they need a loan, who decides, and when they can visit. Four or five questions is plenty on a first call. Each answer should be written to a column your sales team reads before calling back.

Keep the budget question soft. "Roughly what range are you comfortable with?" gets more honest answers than "What is your budget?", and the agent should accept "not sure yet" and move on.

How should the agent handle common objections?

Handle objections by acknowledging them, answering only from approved facts and offering the next small step: a WhatsApp brochure, a call-back from sales or a visit. The agent should never argue, pressure or invent urgency. If an objection needs judgement, such as a price discussion, it books a human call.

How do you book site visits that actually happen?

Book a specific slot on the first call, confirm the pickup or meeting point, send a WhatsApp confirmation straight after, and call again the day before and the morning of the visit. Make rescheduling easy; a moved visit is better than a missed one. Keep visits inside your real site timings.

At Koravi, the site is open Monday to Saturday and pickup is offered from fixed points in town. The agent knows both facts and books only within them, so buyers are never promised a Sunday visit or a doorstep pickup that does not exist. This is the core of site visit booking AI that works: the agent books what operations can deliver.

A visit-booking sequence we recommend:

1. On the first call: offer two specific slots inside site hours; confirm the date, time, number of people and whether they want pickup from one of the fixed points.
2. Within minutes: send a WhatsApp confirmation with the date, time, pickup point and a contact number.
3. The day before: a short reminder call ("Just confirming your visit tomorrow at 11") with an easy option to reschedule.
4. The morning of the visit: a quick confirmation call so the pickup driver and site team know who is coming.
5. If they cancel: offer the next two slots straight away and write the new appointment into the sheet.
6. After the visit: a feedback call the same evening or next day, then a call-back from sales for anyone who wants to proceed.

Reminder and confirmation calls are short, usually under a minute, so they cost little (roughly ₹3–8 each on our pricing). We cover scripts and timing in detail in booking site visits on the first call.

How do you keep an AI agent RERA-safe?

Give the agent only registered, approved facts for each project, and forbid anything else. Under RERA, a promoter must not market an unregistered project and is answerable for false statements in advertisements. So the agent shares approved facts, gives the registration number when asked, and never promises returns, appreciation or dates beyond the registered ones.

This is general information, not legal advice. The provisions that matter most for calling, as summarised on the Andhra Pradesh RERA advertisement guidelines page:

• Section 3: no promoter may advertise, market, book, sell or invite people to buy any plot, apartment or building in a project that is not registered.
• Section 11: the registration number must appear in advertisements, and brochures must be submitted to the authority.
• Section 12: a buyer who relied on an incorrect or false statement and withdraws is entitled to their investment back with interest.

A sales call is part of how you market a project, so we treat the agent's words with the same care as a printed brochure.

The locked-facts lesson

In an early Koravi setup, one agent loaded facts from several projects at once. It mixed up prices and locations between projects, the kind of error that misleads a buyer and creates exactly the risk RERA addresses. Since then, every project gets its own agent and its own locked set of facts. Anything outside those facts becomes a call-back from the project team, never a guess.

What about after-hours calls and event invitations?

Use a separate after-hours agent that never discusses project details and only takes the caller's name, interest and preferred call-back time. Use short, single-language campaigns for event invitations, with an inbound agent ready for people who call back. Both keep buyers looked after without the risk of an agent improvising.

The after-hours front desk

Koravi added an after-hours front-desk line: a separate agent that never gives project details. It takes the caller's name, what they are interested in and a preferred call-back time, and promises a call from the project team. It is a deliberate limit — no prices, no locations, no availability — so nothing said at 10 PM can contradict what sales says the next morning.

Event invitations

For the same developer's ground-breaking (Bhoomi Pooja) event, we ran an invitation campaign in Kannada only, with a separate inbound agent answering people who called back. These calls were short and scripted: invite, give the date and place, confirm interest, and answer a few simple questions. A single language suited the audience, and a dedicated inbound agent meant return calls were answered rather than lost.

How do you hand leads to the sales team and measure results?

Write every call result into the sheet or CRM your team already uses (status, outcome, appointment and last-called time) with a short summary and the recording. Sales calls qualified buyers first. Measure the funnel from enquiry to answered call, qualified lead, booked visit and completed visit, by source and project.

The handover

• Sheet write-back: each call updates status, outcome, appointment and last-called columns, with protection so the agent never overwrites what your team typed.
• Summaries and recordings: sales reads a two-line summary before calling and can play the recording if needed.
• Live monitor: a manager can listen to live calls, whisper to the agent or take over a call that needs a person.
• Daily report: an email summary of yesterday's calls, outcomes and visits booked.

What to measure

For context from our own platform: between May and October 2026 we handled 47,731 calls across several Indian businesses, mostly real estate. Using "lasted 15 seconds or more" as a proxy for a connected call, about 61% of outbound calls crossed that line. Recent sales conversations in October ran around two minutes (median about 122 seconds, on a small sample), which on our plans costs roughly ₹12–16 per conversation. Caveats: campaign types were mixed, duration is only a proxy for a real pickup, and results vary by industry, list quality and time of day.

Common mistakes in real estate AI calling

Most failures we have seen come from setup, not from the technology: shared facts across projects, long scripts, wrong default language, slow first calls and visits booked outside real site hours. Each is cheap to fix before launch and expensive to fix after buyers have heard the wrong thing.

• One agent for several projects. Prices and locations mix. One project, one agent, one locked set of facts.
• Scripts written like brochures. Paragraphs read aloud sound rushed and invite interruptions. Write turns of one or two sentences.
• Calling a day late. A lead from Saturday's ad called on Monday is already talking to someone else.
• Booking visits the site cannot host. Know the days, timings and pickup points, and book only within them.
• No reminder sequence. A visit booked a week out with no confirmation is a visit at risk.
• Letting the agent answer anything. Questions outside locked facts (legal title, loan approvals, offers) go to sales.
• Forgetting follow-up. Many buyers need several touches. Our guide on real estate lead follow-up in India covers what to do after the first call.

If your projects are in North Karnataka, we have live deployments there and can show you how AI calling services in North Karnataka work in Kannada, Hindi and English.

AI calling for real estate works when it does the parts buyers notice (a quick first call, a clear answer, an easy visit) and hands everything else to your team. If you want to hear it on your own project, book a 20-minute demo — we'll call your phone with an agent set up for your business.

Written by the DigitalCallers Team

We build and run AI calling agents for Indian businesses from Gurugram, in Hindi, English and regional languages. Our articles draw on calls we have run ourselves and link to primary sources such as TRAI, DoT and MeitY for anything regulatory. Spotted something out of date? Write to support@digitalcallers.com.