Telegram Bot for a Real Estate Agency: Matching and Requests
A real estate agency loses leads not because of weak advertising, but because a manager fails to reply within the first fifteen minutes. A Telegram bot takes the request around the clock, asks clarifying questions and shows real properties with photos before anyone picks up the phone. Below is how this works in practice, how much time it takes and what the business actually gets.
What a Telegram bot for real estate does
The bot works as a first line of contact: it is available at any hour, replies instantly and does not depend on how many managers are on shift today. The client enters the chat from an ad, a listing or a QR code on a banner near the property, taps Start and immediately lands in a dialogue that asks the essentials: budget, district, number of rooms and deal type - rent or purchase.
The whole property database is pulled from your CRM or spreadsheet. The bot does not simply forward a link to the website, it shows full cards: photos, area, floor, price, a short description and buttons for booking a viewing or showing similar options. The client sees concrete offers in 30-40 seconds instead of waiting half a day for a reply while messaging three other agencies in parallel.
For the agency this means no contact is lost. Everyone who wrote to the bot is already in the database with a phone number, search criteria and a history of viewed cards. Even if the person does not buy now, in three months you know exactly what to offer and do not start the conversation from scratch. This matters most on expensive properties, where the deal cycle stretches over half a year.
A filter questionnaire and property matching
The questionnaire is the heart of the bot. It is kept short: 4-6 questions with buttons, without free text input wherever a choice will do. Budget is set in ranges, district by list or map, room count by buttons from one to four plus. The less a person has to type on a phone, the higher the share of those who reach the end of the form instead of dropping off at question three.
After the answers the bot queries the database and returns the 3-5 most relevant options. Showing more makes no sense: the person gets lost in the list and closes the chat. If there are no exact matches, the bot widens the criteria itself - for example, adds a neighbouring district or raises the upper budget limit by 10 percent - and honestly warns that these are close, not exact, matches.
Filters are easily extended to the agency's specifics: new build or resale, renovation state, not the first or last floor, pets allowed, parking, move-in readiness, documents ready. Every extra filter narrows the selection, so only the two or three that genuinely affect the decision go into the form, while the rest are left for the agent to clarify in conversation.
Viewing request intake and handoff to an agent
Viewing request intake is where the bot brings in the most money. Under each property card there is a booking button: the client picks a convenient day and time slot, the bot checks the agent's free slots and locks the meeting into the calendar. The agent gets a message with the address, contact and client criteria - no correspondence across three chats and no digging for a phone number in old messages.
A day before the viewing the bot sends a reminder with the address, a landmark and Confirm or Reschedule buttons. It is a simple thing, but it removes a large share of wasted trips where the agent drives across the city and the client has simply changed their mind and forgotten to say so. People happily reschedule in one tap, but they almost never call to apologise.
When the conversation goes beyond the script - questions about mortgages, price negotiation, legal status or the condition of utilities - the bot hands the dialogue to a live agent. The handoff happens inside the same chat: the client does not have to go anywhere, and the agent sees the full prior history, the chosen filters and the viewed cards, so nothing already answered gets asked twice.
New listing notifications and bringing clients back
Most buyers are not ready to close in the first week. So the bot remembers the criteria and sends new listing notifications: as soon as a property matching a specific client's filters is added to the database, they receive a card with photos and price on the day it is published. This works better than any mass mailing, because the message is personal and concerns exactly that district and budget.
Frequency should be capped - no more than one or two messages a week, with a visible Pause matching button. Aggressive mailings in Telegram quickly end in the bot being blocked, and you lose the channel forever along with the whole dialogue history. A sensible limit, by contrast, keeps the base alive for months and turns the bot into a source of repeat enquiries.
The bot also reports price drops on properties the client already viewed, and status changes: reserved, sold, available again. These triggers bring back people who seemed to have dropped out of the funnel - a five percent price cut often turns out to be a stronger argument than ten manager calls asking whether they are still looking.
Lead qualification and setting the bot up
Lead qualification happens invisibly for the client. The bot tracks whether a realistic budget was given, whether there is urgency, whether a mortgage is needed, how many cards the person opened and whether they booked a viewing. Based on that the lead gets a score and the manager sees in the CRM who to call first. Instead of dialling everyone, the agent works with a dozen hot contacts a day while cold ones mature on automatic notifications at no human cost.
Setup takes from two to four weeks, depending on the format the property database is stored in and whether two-way CRM integration is needed. The longest part is usually not the programming but cleaning up the data: photos of varying size and quality, empty area fields, duplicates of the same property from different agents, districts spelled in five different ways.
It is worth starting with a minimal scenario: questionnaire, matching, viewing booking. Everything else - notifications, scoring, analytics by traffic source - is added later using data from the first months, once it is clear where people actually leave the dialogue. At Devlly we build such bots for a specific agency: with your property database, your filters and your CRM, rather than from a universal template.