Real estate

How to reply to every property lead in minutes with ChatGPT

Set up one ChatGPT Project with your inventory and rules, then answer every portal and WhatsApp enquiry in about three minutes.

Published 25 Oct 2026 · 6 min read
← Back to blog
Tools used
ChatGPTClaudeWhatsAppGoogle Sheets

Most property enquiries arrive at odd hours, from 99acres, MagicBricks, Housing.com or the WhatsApp number printed on a hoarding. A reply that takes four hours often means the buyer has already spoken to two other brokers. This guide shows a sales executive in a Mumbai, Pune or Thane office how to build one ChatGPT Project that drafts a correct, warm reply to each enquiry, in English or Hinglish, using only the facts your team has approved.

What you need

  • A ChatGPT account where you can create Projects. OpenAI lists Projects on free and paid plans, but file limits depend on your plan, so check yours first.
  • Your current inventory sheet: project name, configuration (2 BHK, 3 BHK), carpet area, price range in ₹, possession month and RERA registration number.
  • Your site visit timings and the name of the person who takes visits.
  • Five to ten real enquiries from the last month, with the replies that worked best.

Step 1: Create a Project for your inventoryChatGPT

Open ChatGPT, go to Projects in the sidebar and create a new project. Name it after the development or the team, for example "Pune Sales Enquiries". OpenAI's help article describes a project as a place that groups chats, reference files and custom instructions, so every chat you start inside it sees the same material.

Step 2: Upload the facts, not the brochureGoogle Sheets

Upload your inventory sheet as a CSV or PDF, plus a short text file with site visit timings, the booking steps and the documents a buyer needs. Keep the files short and accurate. If a price is not in the file, the reply should not invent it. Update the files every Monday, because an old price is the most common mistake in this kind of setup.

Step 3: Write the project instructionsChatGPT

In Project settings, add instructions. OpenAI's help article explains that project instructions apply only inside that project and take priority over your general custom instructions. Paste something close to this:

Prompt

You are the enquiry assistant for [Company name], a real estate developer in [city]. Use only the facts in the uploaded inventory file and FAQ. If a price, possession date, offer or availability is not in the file, do not guess. Instead write "I will confirm this with our sales manager and come back to you by [time]". Keep replies under 80 words, warm and direct. Match the buyer's language, English or Hinglish. Never promise returns, appreciation or guaranteed rent. End with one question that moves the buyer toward a site visit.

Step 4: Draft a reply to one enquiryWhatsApp

Copy the enquiry exactly as it arrived. Add the source and the budget if the buyer mentioned it.

Prompt

Here is a new enquiry from [portal or WhatsApp]: "[paste enquiry]". Budget mentioned: [budget or not stated]. Draft one reply I can send on WhatsApp. Give me the reply first, then a two-line note for me on what the buyer seems to want.

Read the draft against the inventory file before you send it. Check the price, the configuration and the visit slot. This takes about a minute, and it is the step that stops a wrong quote from reaching a buyer.

Step 5: Sort a batch of enquiries

At the end of a busy day, paste the list of ten or twenty enquiries.

Prompt

Below are today's enquiries, one per line, with the time and the source. Sort them into three groups: hot (budget and timeline given, wants a visit), warm (interested, no timeline yet) and cold (price check or wrong location). For each hot lead, write a one-line follow-up for the sales manager. Enquiries: [paste list]

The sorting is only as good as the notes you give it, so write the budget into the enquiry whenever a buyer mentions one.

Step 6: Save your best replies as WhatsApp quick repliesWhatsApp

The WhatsApp Business app lets you save messages as quick replies. Pick the five replies your team uses most, such as the visit confirmation, the brochure request and the price follow-up. Save them with short shortcut names. Anyone on the team can then type a slash and the shortcut to insert the message. Check the price and the date in each one before sending, because a saved reply with an old price causes more trouble than a slow reply.

Step 7: Write the follow-up after a site visitChatGPTWhatsApp

Give ChatGPT the visit notes in plain words and ask for the follow-up.

Prompt

The buyer visited [project] today. Their notes: [liked the balcony, worried about the metro distance, wants a 3 BHK under ₹1.4 crore]. Write a WhatsApp follow-up under 70 words. Mention one fact from the inventory file that answers their worry. Do not add any new offer or discount.

Tips

  • Keep a short list of phrases your company never uses, such as "guaranteed returns" or "assured appreciation", and put it in the project instructions.
  • Ask for Hinglish when the buyer writes in Hinglish. A reply in the buyer's own words usually gets a faster answer.
  • Review every reply that quotes a number, a date or a RERA detail. Those details must match the registration.
  • Ask the manager to approve any new claim before it goes into the files.

Keep it safe

Do not paste buyers' Aadhaar numbers, PAN, bank details or loan papers into a personal ChatGPT account. Enquiry text with names and phone numbers belongs in your CRM. Use a business or team account, check its data settings before the first upload, and keep the uploaded files limited to the inventory sheet and the FAQ.

Sources

Cut costs. Get hours back.

In 15 minutes we show where AI saves your team the most time and money.

Get my savings plan