DOC // AIQ-INMO-001-ENREV / A
●SOLUTIONSHEET 09 / REAL ESTATE
§ Solution · Argentine real estate
Your agency answers inquiries, assigns visits and reports — without anyone doing it/
"The client writes at 10pm. At 10pm, the assistant answers."
An AI agent connected to WhatsApp, Telegram and email that classifies every inquiry, searches properties across your own listings and external portals, coordinates your agent team, and sends you an executive summary every day at 8pm. No software to install. No tools to switch.
< 30 sec
Response time
Before: hours. With the assistant: seconds. Any time of day.
24/7
Continuous coverage
60% of real estate inquiries arrive outside office hours.
3 channels
Unified
WhatsApp, Telegram and email in a single system. History persists across channels.
14 days
To a running system
From the first meeting to the first client served automatically.
§01 · Who it's for
Does your agency have these problems?
This solution applies directly if you manage properties in Argentina, have between 1 and 20 agents, and recognize your business in any of these scenarios:
Signs this solution is for you
- You get inquiries at night or on weekends and there's no one to answer — the lead calls another agency
- Your agents spend hours answering the same things over and over: price, size, HOA fees, availability
- You coordinate visits over personal WhatsApp and end up with double bookings, no-shows or schedule conflicts
- You have more than 5 agents and assigning properties by zone is a manual, error-prone process
- You don't know how many inquiries you got yesterday, which channel they came from, or which property drew the most interest
- You use a shared spreadsheet nobody updates in time, and leads get lost between chats
- You want to offer search across Zonaprop, Argenprop and MercadoLibre without your team opening each portal
- As the owner, you want to know how the day went without having to ask anyone
§02 · What it does
Five operational capabilities
[ 01 ]
Searches your own listings
The client asks, and the assistant queries your property Google Sheet in real time. It filters by type, area, price, rooms and status. It answers with the exact data from the listing — no invented facts, no improvising. If the data isn't there, it says so and offers direct contact.
[ 02 ]
Searches external portals
When the client asks for something you don't have in your own listings, the assistant runs parallel searches on Zonaprop, Argenprop and MercadoLibre. Gemini parses the results and replies with the most relevant options: price, area and a direct link to the original listing.
[ 03 ]
Schedules and assigns visits
The assistant detects visit intent, identifies the agent available for that area and notifies them via Telegram. The lead is logged in Sheets with status "Visit requested." The client gets confirmation on the same channel where they asked.
[ 04 ]
Answers technical questions
HOA fees, covered square footage, building age, floor, parking — available without bothering any agent. The assistant reads the full property listing and answers precisely. Average response time: under 15 seconds.
[ 05 ]
Reports to the owner every day
Every day at 8pm, Gemini generates an automatic executive summary: the day's inquiries by channel, most-viewed properties, scheduled visits, and leads needing urgent follow-up. Delivered by email and Telegram. No extra app to open.
[ ● ]
Memory across channels
If a client starts on Telegram on Monday, continues by email on Tuesday and wraps up on WhatsApp on Wednesday, the assistant remembers the whole conversation. History persists in a Data Store and is retrieved on every interaction, regardless of channel.
§03 · Architecture
How it works under the hood
The system is built on 5 Make.com scenarios that communicate with each other. Every incoming message triggers the full flow in under 30 seconds.
01
Channel receivers
Three separate scenarios listen on WhatsApp (Twilio webhook), Telegram (WatchUpdates) and Gmail (IMAP). Each one normalizes the incoming message into the same canonical format and passes it to the Central Agent.
02
Central Agent
Reads the conversation history from the Data Store. Gemini 2.5 Flash classifies the message's intent as: LISTING_QUERY, EXTERNAL_SEARCH, SCHEDULE_VISIT, PROPERTY_QUESTION or ESCALATE. It then runs the matching branch.
03
Response dispatcher
Receives the generated reply and sends it to the originating channel: Telegram, WhatsApp (Twilio) or Gmail. The lead is logged in Sheets. The agent gets their notification. History is updated in the Data Store.
WhatsApp Business (Twilio)
Telegram Bot
Gmail (IMAP / OAuth)
WhatsApp Meta API roadmap
§04 · External portals
Zonaprop, Argenprop and MercadoLibre — in real time
When a client asks about a property you don't have listed, the assistant doesn't say "we don't have that." It searches the three major portals and presents the most relevant options with price, area and a direct link to the listing.
The full process — from the client's message to a reply with results — takes under 20 seconds.
Sample conversation
- Client on Telegram: "Find me houses for sale in Tigre up to USD 280,000"
- Assistant: makes 3 parallel requests to Zonaprop, Argenprop and MercadoLibre
- Gemini: parses the HTML and selects the 3–5 most relevant options
- Assistant: "Found 4 options in Tigre: → 4-room house · USD 265k · Boulogne · View listing"
ZONAPROP
Leading portal for residential rentals and sales in Argentina. Highest volume of listings in Buenos Aires and Greater Buenos Aires.
ARGENPROP
Strong presence in commercial properties and inland provinces. Complements residential search well.
MERCADOLIBRE INMUEBLES
Highest traffic in LATAM. Captures demand from buyers who don't use specialized portals.
* If the portals block Make.com IPs, ScraperAPI is added as a proxy (~US$ 49/month, optional).
§05 · Impact
Before and after, in concrete operations
Inquiry at 10pm → no reply until the next day. The lead calls another agency.
Reply in under 30 seconds, any time. The lead gets options before calling anyone else.
Overloaded agents answering price, size and HOA fees instead of closing deals.
Agents receive pre-qualified leads with full history. They focus on negotiating and closing.
Double bookings and no-shows because visit coordination is manual, over personal WhatsApp.
The right agent for the area gets the notification on Telegram. No errors, no duplicates.
Shared spreadsheet nobody updates. Leads lost between chats. Monday meeting just to catch up.
Every inquiry logs itself in Sheets. Pipeline visible in real time. No follow-up meetings needed.
The owner doesn't know how the day went until they ask. Decisions made on 3-day-old data.
Automatic executive summary at 8pm with inquiries, hot properties and the day's visits.
§06 · Technology
Technical stack
| Tool |
Role in the solution |
Approx. cost |
| Make.com |
Orchestrates the 5 scenarios. Webhook trigger, processing, routing and delivery. |
~ US$ 29/mes |
| Google Gemini 2.5 |
Flash for intent classification, Pro for drafting replies and parsing portals. |
~ US$ 10–30/mes |
| Google Sheets |
Database of properties, agents and leads. Editable directly by the team, no code. |
Included with Google Workspace |
| Twilio |
WhatsApp Business API. Dedicated agency number. Receives and sends messages. |
~ US$ 15/mes + mensajes |
| Telegram Bot |
Channel for clients and internal notifications to the agent team. Free. |
Free |
| Gmail |
Inbound and outbound email channel. Daily reports to the owner. |
Included with Google Workspace |
| Make Data Store |
Persists conversation history per contact across channels and sessions. |
Included with Make |
* ScraperAPI (~US$ 49/month) optional if the portals block Make's IPs in production.
Total opex: ~ US$ 100–120 / month
Make.com
Google Gemini 2.5
Google Sheets
Twilio
Telegram
Gmail
No lock-in
§07 · Requirements
What you need on your end
[ REQ ]
WhatsApp Business number
A dedicated number for the agency. Can be a new SIM. Twilio manages WhatsApp Business API activation — the process takes 2–3 business days.
[ REQ ]
Gmail account
The agency's email for handling inquiries and receiving the daily report. If you already have Google Workspace, that account is used directly.
[ REQ ]
Property listings
A basic list in any format: address, price, type, rooms and status. If you already have a spreadsheet, we import it into the solution's Sheets on day 1 of implementation.
[ REQ ]
Agent list
Name, coverage area and Telegram contact for each agent. With this information the system assigns visits automatically by geographic area.
[ REQ ]
Telegram bot
We create it with @BotFather in 5 minutes during setup. No paid account required — Telegram is free for both the client channel and the internal team.
[ OPT ]
Previous inquiry history
If you have a CRM or spreadsheet with past leads, we can import them into Sheets for historical context from day one. Not required to get started.
§08 · Implementation
From zero to a running system in 14 days
STEP 01 · DAY 1–2
Discovery
We map your current listings, active channels and agent workflow. We set up the Google Sheet for properties, agents and leads with the solution's schema. If you already have data in a spreadsheet, we import it.
STEP 02 · DAY 3–7
Configuration
Setup of Twilio WhatsApp Business, Telegram bot and Gmail connection. Import and configuration of the 5 Make scenarios with your real connection IDs. You don't write any code — we connect the pieces.
STEP 03 · DAY 8–12
Pilot
The system handles real inquiries from your listings. We tune Gemini's prompts, classification rules and agent filters using live data from your business. Results visible from day one.
STEP 04 · DAY 13–14
Handoff
Team training: how to read the leads Sheet, how to update listings and how to interpret the daily reports. We leave the system documented and ready for you to run without depending on anyone.
SUPPORT · DAY 14+
30 days included
30 days of post-handoff support at no extra cost. Prompt fixes, flow adjustments and response to any integration issue. No software installed on your end — everything runs in the cloud.
Implementation cost
US$ X.XXX
To be agreed based on listing volume and required channels. Final price set in the first meeting.
§09 · FAQ
What always comes up before deciding
Does the client know they're talking to an AI?
Not necessarily. The assistant replies under the agency's name, in natural language. If the client asks to speak with a person, the system escalates immediately to the agent and notifies them on Telegram with the full conversation context.
What if the AI gives incorrect information?
The assistant answers using the data you load into Sheets — it doesn't invent or hallucinate. If the client asks something not on the listing, the system says so and offers to connect them with an agent. You can audit every conversation in the Data Store history.
Are my data and my clients' data secure?
The data lives in your Google Sheets, in your own Google account — not ours. Gemini processes the message text but doesn't store it or use it to train models. Make.com is GDPR and ISO 27001 compliant. You can audit what's stored at any time.
Does this replace my agents?
No. The assistant handles repetitive inquiries and filters leads. Agents receive pre-qualified prospects with full history and focus on negotiating and closing. With less time on WhatsApp, they're more productive.
Can I update the listings myself?
Yes, directly in the Google Sheet. Add a row, change a property's status to "Reserved" or update the price — the assistant uses the updated data on the next inquiry, with no code to touch and no one to call.
Want to see it in action?
Book a call and we'll show you the system running with real data from a real estate agency: the chat, the visit assignment, the reports. No generic slides.
END OF SHEET 09
ai-q.fit · real estate
© 2026 · Dr. Federico Andino