An AI sales manager for real estate developers

What is an AI sales manager for a developer?

An AI sales manager is a chat assistant that takes the first line of a developer’s sales work. It answers every buyer who writes, within seconds and at any hour, in the language the buyer wrote in. It knows your apartments, prices and payment plans, and it answers from that list only. It asks the questions a sales manager would ask, books a call or a viewing, and puts the lead and the whole conversation into your CRM. From there, one of your sales managers takes the deal.

In other words, it is a chatbot for a developer’s sales team that talks like a sales manager instead of offering buttons. That is the whole job. It doesn’t negotiate, it doesn’t give discounts and it doesn’t sign anyone up.

We have not built one for a developer yet, so there is no developer case or result on our site. What we can show are the parts, because we have built each of them before in other projects: answering only from given data, scoring leads with AI and drafting replies next to a conversation. We come back to each below.

Why does the first reply have to be fast?

Because an online lead goes cold within the first hour, not the first day. In a study published in Harvard Business Review in 2011, firms that tried to reach an online lead within an hour were nearly seven times as likely to qualify it, which the authors defined as a meaningful conversation with a key decision maker, as firms that tried even an hour later. Against firms that waited 24 hours or more, the gap was over 60 times. The same authors audited 2,241 U.S. companies: 37% answered a web inquiry within an hour, and 23% never answered at all.

Those are American figures from many industries, not real estate in Georgia, but the logic holds for a buyer of a flat. Someone looking at new builds from abroad may write to several developers in one evening, after your office has closed. The developer who answers first, with a real price and a real payment plan, is the one most likely to get the call.

A person can’t sit by the chat at night and at weekends. An assistant can, and it answers the tenth question about the same studio as patiently as the first.

Who writes to a developer in Batumi, and in which language?

Often a buyer from abroad, writing in their own language. In Galt & Taggart’s survey of Batumi developers, foreign buyers accounted for 52% of apartment sales in 2025. Buyers from Israel made up 13% of sales, buyers from the EU another 13%, buyers from Ukraine, Russia and Belarus 11%, and buyers from Turkey and from Central Asia 3% each. The survey covered developers with over 30 ongoing projects, about 40% of the city’s new-build market. Across new and resale homes together, Batumi recorded 17,478 apartment sales that year.

Buyers from that many countries don’t share one language, and few sales teams speak all of theirs. The AI sales manager answers in the language the buyer writes in and hands the conversation to your team with a short summary, so a manager who speaks English or Russian can still pick up a lead that began in Hebrew or Turkish.

How does it avoid inventing a price or a free apartment?

It answers from your data and nothing else. We load your projects, the apartment list, prices and payment plans, and the assistant may quote only what is in them. If a buyer asks about a unit that isn’t on the list, a discount, or a date the list doesn’t give, it says a manager will confirm and passes the question on. A made-up price in a chat is worse than no answer: the buyer remembers it, and your manager has to take it back.

We learned this rule on TenderLens, an AI tool that reads tender documents and fills a brief for each tender. Early on, the client caught a brief that turned a 5% bid security into “5 to 30%”. Since then the AI there may not fill in values the documents don’t give or guess market prices, and the same documents get the same brief every time.

For a developer, the same rule rests on one thing on your side: the list has to be current. A unit sold this morning must leave the list the assistant reads, whether that list is a spreadsheet your team already keeps or your CRM. Where the list lives and who updates it is one of the first things we settle in the scope.

What does it ask, and what reaches your sales team?

It asks what a good sales manager finds out in the first five minutes, and your team gets the answers, the conversation and a booked call or viewing. The questions:

• the budget;
• the purpose: to live in or to rent out;
• the timing: buying now, this year, or just looking;
• how they want to pay: in installments or in full;
• the size and the project they are looking at.

They come one at a time, inside the conversation, not as a form. The purpose matters a lot in Batumi: Galt & Taggart calls rental yield the key demand driver for investment-led projects, and a buyer who plans to rent out asks different questions from one who plans to live there.

Then the assistant offers a call or a viewing at a free time and books it. The lead lands in your CRM with the answers, the conversation and a note on how ready the buyer seems. We built that kind of note before in TGtrigger, a former product of ours that scored leads with AI: it labelled each one hot, warm, cold or spam and added a one-line reason. The view next to the conversation comes from Pipey, our own beta CRM for people who sell in messengers, now in our archive: there an AI panel drafts a reply and sums up the thread beside each chat, and a person decides what to send.

Here is how the work splits between the assistant and your team:

If you don’t have a CRM yet, the leads can go to your sales team’s chat in Telegram, or into a simple CRM built around your sales.

Where does it work: website, Telegram or WhatsApp?

In any of the three, and it knows the same things in each. On your website it is a chat window on the project pages. In Telegram it is a bot that buyers write to directly. In WhatsApp it works through Meta’s official WhatsApp Business Platform, under your company’s business account on Meta, not through the WhatsApp app on a sales manager’s phone.

WhatsApp has one rule worth knowing before you start. When a buyer writes to you, a 24-hour customer service window opens, and inside it the assistant can reply with ordinary messages. Once the window closes, you can only send templates Meta has approved in advance, and those count towards a limit: Meta’s documentation sets it at 250 people per 24 hours for a new business portfolio, and verifying your business with Meta is one way to raise it. We set the account up together with you.

Start with one channel, the one where most of your buyers already write. A second channel later uses the same data and the same rules.

What does it cost, and how do we start?

AI tools start from $500 and take from 2 weeks. We don’t publish a separate price for an AI sales manager: it depends on your setup, and you get the exact figure in a written scope after a free call. What moves it:

• how many projects and apartments the assistant has to know;
• the channels: website chat, Telegram, WhatsApp;
• the languages your buyers write in;
• where your price list and availability live, and how they reach the assistant;
• your CRM, and what should land in it.

You can start small: one project and one channel. That is enough to see what buyers ask at night, how many leads come in and where your price list needs work, before you add the next project.

It runs like our other projects. After the call you get a written scope with a fixed price, which changes only if the scope does. You try a working prototype on your own price list before the bulk of the work starts. The person who scopes the project builds it. We work remotely from Batumi, Georgia, in English or Russian. Our article on where AI saves time in a small business shows the same approach in our other AI projects.

Questions

Sources

1. Harvard Business Review: The Short Life of Online Sales Leads (James B. Oldroyd, Kristina McElheran, David Elkington, March 2011)
2. Galt & Taggart: Batumi Residential Real Estate, 2025 Review and 2026 Outlook (6 March 2026)
3. Meta for Developers: WhatsApp messaging limits
4. Meta for Developers: Sending messages on WhatsApp, customer service windows

Alexander ZabrodinCEO · Founder

ZabSoft is a software company in Batumi. The person who scopes your project is the person who builds it.