An AI assistant that greets every WhatsApp lead within seconds, clarifies their needs across four criteria, finds the match by meaning in your portfolio, and books the viewing. Your agents only talk to ready buyers.
Want to start right away? Join the early-access program →
Early-access program
Early-access offices shape the product with us and get dedicated setup, training and process support. Offices that complete the program keep a permanent advantage.
Free setup + portfolio import
We import your Sheets/Excel in a day; zero manual work.
Dedicated onboarding call
One-on-one with you, tuned to your office's real scenarios.
Direct founder support
Your questions go straight to the founder on WhatsApp; no support tickets.
Founding-office advantage
Early-access offices keep working with a permanent discount.
How it works
You don't manage the steps in between. WhatsApp → needs analysis → portfolio matching → appointment chain — completed without an agent stepping in.
Customer writes
A customer coming from an ad, a listing or a referral sends a single WhatsApp message. No app download, no forms.
AI clarifies the need
The AI assistant collects the 4 mandatory criteria: transaction type, area, room count, budget. It understands the area down to the neighbourhood; a 10% budget tolerance is applied automatically.
Finds the fit in the portfolio
The customer's intent is matched against listing descriptions by semantic similarity; budget and transaction filters work alongside. At most 5 listings are sent, ranked by fit.
The agent connects
When the customer likes a listing, a viewing is scheduled. The relevant agent gets the details; the others are kept informed.
Customer writes
A customer coming from an ad, a listing or a referral sends a single WhatsApp message. No app download, no forms.
AI clarifies the need
The AI assistant collects the 4 mandatory criteria: transaction type, area, room count, budget. It understands the area down to the neighbourhood; a 10% budget tolerance is applied automatically.
Finds the fit in the portfolio
The customer's intent is matched against listing descriptions by semantic similarity; budget and transaction filters work alongside. At most 5 listings are sent, ranked by fit.
The agent connects
When the customer likes a listing, a viewing is scheduled. The relevant agent gets the details; the others are kept informed.
System promise · measured together with early-access offices
< 30 s
First reply
Needs analysis starts the moment the customer writes
24/7
Automatic greeting
Midnight, weekends, holidays — makes no difference
4 criteria
Mandatory analysis
Transaction type, area, rooms, budget — complete
10%
Budget tolerance
Adjustable headroom that leaves space to negotiate
KVKK
Compliant infrastructure
Hosted in Türkiye, data ownership stays with you
3 steps
Setup
Portfolio import + tuning + live test
Live demo
In the box below you're talking to a real AI search. The system understands your criteria and returns matching listings from the portfolio. It uses the same AI as the real flow — it just doesn't send WhatsApp messages, and there's a monthly usage cap.
Realtiq Assistant · Live
connected to real AI
The Live Demo connects to real AI. No appointments or WhatsApp messages are sent.
Realtiq Assistant · Quick Demo
scripted flow
A pre-built scenario — no API calls, quota or verification.
Quick demo
A scripted scenario. No connection, quota or verification — it simply shows how the system designs the customer journey. Ideal for a quick look at how knowing the customer → portfolio → appointment chain plays out.
Features
Realtiq is not a message autoresponder. Knowing the customer, matching by meaning, booking viewings, reporting and security filters come together in a single system.
Knowing the customer
While chatting with the customer, the AI assistant extracts the transaction type, area, room count and budget step by step. If any field is missing, no listings are sent; the question-answer loop completes it first.
Matching by meaning
The intent the customer writes is matched against listing descriptions by semantic similarity; the transaction-type and budget filters always run alongside. At most 5 listings are sent, ranked by fit.
Region intelligence
The assistant knows Türkiye's official geography — 81 provinces, 973 districts, over 73 thousand neighbourhoods. When there's no exact listing in the requested area, genuinely adjacent neighbourhoods are suggested; far-away listings never make the list.
Appointment chain
The moment the customer says "I'd like to see it", an appointment record opens. The agent of the best-fitting listing gets the details; other matching agents are kept informed. The customer stays in a single conversation.
Security shield
Every incoming message passes through a protection shield before reaching the assistant. Abusive content, prompt injection (messages designed to mislead the system) and media files are blocked; message floods hit the shield too. These events are collected in the Error Log and never pollute your lead pipeline.
Management dashboard
One dashboard: Overview, Customers, Appointments, Portfolio, Reports and the Error Log. No data is ever shared with another office; what we collect from you stays with you alone.
Integration
Add a listing or update a price, and it appears in the system within 30 minutes. The system reprocesses only what changed and leaves the rest untouched. You don't have to change a single habit.
Region intelligence
The assistant knows Turkish geography down to the neighbourhood: which district every neighbourhood belongs to, and which neighbourhoods are genuinely adjacent — straight from official records. So every suggested listing is either exactly where you asked, or truly close by. A property 100 kilometres away is never sold to you as "nearby".
81
provinces
973
districts
73,496
neighbourhoods
1.27M
streets
aligned with official address and statistics registries
It knows true adjacency
Five distinct tiers. When there's no exact match, genuinely adjacent options come first; a listing across the city never makes the list.
It isn't fooled by duplicate names
There's a Bahçelievler in dozens of cities. The system pins the search to the right one — and when it can't be sure which you mean, it asks instead of guessing.
It tolerates spelling variations
With or without Turkish characters — what the customer writes is matched to the official registry name.
It gives honest results
If there's no listing exactly where you asked, it says so openly and offers the nearest alternatives labelled as "adjacent". It never markets far as near.
Illustrative flow — out-of-area listings are dropped before the list is sent.
Management dashboard
Managers see every customer in the office; agents see their own. Role-based permissions, real process steps, a single source of data.
Total Leads
142
+24%
Appointments Booked
38
+12%
Active Portfolio
248
−4 listings
System Events
17
rate limit · media · guardrail
Sample view — the real dashboard fills with your office's data.
Sound familiar?
The vast majority of home buyers in Türkiye make first contact on WhatsApp. Only a small share of offices manage to reply within minutes. The difference is a lost viewing.
If a WhatsApp message goes unanswered for 30 minutes, nearly half of buyers write to another office. Evening and weekend messages are left waiting until the next day.
The AI assistant steps in the moment a message arrives; 24/7, whether it's the middle of the night or the weekend, it starts collecting what the customer is looking for.
When an agent juggles 30 chats at once, all of them stall. Stuck in the "What's your budget, how many rooms?" loop, the customer gets bored and goes quiet.
The assistant won't move to the portfolio until all 4 mandatory details (transaction type, area, rooms, budget) are collected. Each detail is asked one at a time, one per message.
With hundreds of active listings, it's unclear which one fits which customer. Some listings haven't been shown for months — and no report notices.
Every listing is stored by its meaning, and matching listings arrive ranked by fit. The CRM Portfolio page lists properties unmatched for 30+ days separately.
Security & data
Realtiq is designed so real-estate offices keep full control over their own data. What we collect from you is never served to anyone else.
Customer conversations, lead details and appointments belong to you alone. They're not shared with other offices, not sent to third-party ad networks, and not used for model training.
Data collection, storage and deletion follow the requirements of Türkiye's data-protection law (KVKK). Deleting or exporting data at a customer's request takes a few clicks, with a full audit trail.
Database and application servers are hosted in Türkiye or the EU region. For offices particular about data ownership, a self-hosted deployment on your own server is also available.
All traffic travels over encrypted connections (TLS); access is protected with role-based authorization (admin / office manager / agent).
Insults, flirting, prompt injection (messages designed to mislead the system) and media files are blocked before they reach the assistant. Message floods hit the protection shield. Every block is logged in the CRM Error Log.
The Live Demo on realtiq.io runs in an environment separate from the real WhatsApp flow. Demo messages are never written to your real customer records; a monthly usage cap and Cloudflare Turnstile bot protection keep it safe.
Frequently asked questions
Got another question? Write it in the note field of the early-access form — we'll call you.
Realtiq is a WhatsApp-based AI assistant for real-estate offices. It chats with the customer to clarify their needs across four criteria (transaction type, area, room count, budget), uses AI to find the listings in your portfolio that match by meaning, and books the viewing. It ships as a single system together with a CRM dashboard.
Realtiq works exclusively on WhatsApp. Web chat and SMS are not on the roadmap right now, because in Türkiye the vast majority of home buyers make first contact on WhatsApp. The Live Demo on realtiq.io is the one exception — a sandbox that exists purely so you can try the system.
A greeting comes back within seconds. The system applies flood protection to bursts of messages sent in a very short time; every customer below that threshold is greeted instantly — night, weekend or holiday makes no difference.
Your Google Sheets works as the single source of truth. The system reads your sheet every 30 minutes, reprocesses only the listings that changed, and automatically closes the ones taken off the market. Add a listing or change a price, and it shows up in the system within half an hour. You never have to leave Sheets.
A default 10% tolerance is applied per customer (adjustable per office). When a customer says their budget is 4M, the system may also show listings around 4.4M — stating the price difference transparently. The tolerance also feeds into the lower budget bound of the matching search.
Yes — the assistant knows Türkiye's official geography (81 provinces, 973 districts, over 73 thousand neighbourhoods) and which neighbourhoods are genuinely adjacent to each other. When there's no exact listing in the requested area, physically adjacent neighbourhoods are suggested first; far-away listings never make the list. Duplicate neighbourhood names (e.g. Bahçelievler) are pinned to the right city, and when it can't be sure, it asks the customer. When a nearby alternative is offered, that is stated explicitly.
It is never shared. All conversations, lead details and appointments belong to you alone. Nothing is sent to third-party ad networks or used for model training. In line with Türkiye's data-protection law (KVKK), deletion and export requests are handled with a full audit trail.
Every incoming message passes through a protection shield before it reaches the assistant. Insults, flirting, prompt injection (messages designed to mislead the system) and media files (image/video/audio/document) are blocked. These events never pollute your lead pipeline; they're collected in a separate Error Log.
The lead agent is the one whose listing fits best, and they contact the customer. Agents whose listings also matched are kept informed, but the customer stays in a single conversation. That keeps both the customer experience and in-office coordination simple.
The Live Demo connects to real AI, runs the same semantic search as the production system, and is protected by a monthly usage cap plus Cloudflare Turnstile bot protection. The Quick Demo (scripted) walks through a pre-built scenario step by step — no connection, quota or verification. The first lets you try the system; the second lets you watch the flow quickly.
Realtiq is a CRM in its own right: customers, conversations, appointments, portfolio, reports and the error log all live in one dashboard. If you already use a CRM, we design the transition together with you during early access.
From zero to live in 3 business days. Day one: portfolio import; day two: tuning to how your office speaks; day three: live testing with your agents. You can leave whenever you like, and your data is exported on the way out.
In a 15-minute screen-share call, we run the customer journey live with sample listings from your own portfolio. You see the real system, not a slide deck. The decision stays with you.