How does your real-estate project lead AI answers?

AI real-estate marketing, in the sense this page covers, means an AI engine citing your project's page when a customer asks it about a district, a price or a size. It does not mean writing your ads with an AI tool. The first puts you inside the answer the buyer or tenant reads; the second only saves you writing time.

An AI engine is a program that reads pages on the web, writes an answer in its own words, and attaches links to the sources it relied on, such as ChatGPT, Gemini and Google AI Overviews. The link it places beside a sentence in its answer is the citation, and it is the route a customer takes to your site. The eight engines we measure are described on the AI engines we measure in Arabic.

Someone looking for an apartment to rent or an off-plan villa used to open ten listings and compare them by hand; today they type one question with every condition in it and read a paragraph that does the comparing for them. Whoever is named in that paragraph makes the call list; whoever is not never learns they dropped off it.

The work that gets you into that paragraph is optimising for AI answers (GEO). This page applies it to property, for developers and brokers alike: what your customer asks, who is cited instead of you, the page to write, and how to measure the result.

Four things before anything else: the district, the price, the size, and the distance to the metro or a main road. Then a condition of their own: the number of rooms, a school nearby, a down payment they can manage, or a handover date.

The old search keyword was "apartments for rent Riyadh"; the question today is a full sentence carrying the district, the rooms, the budget and the distance to the station, and the engine looks for a page that answers all of those conditions together. A page that says "luxury apartments in Riyadh's finest districts" answers none of them.

We sort property questions into six families, and each needs an answer written on your page, not saved for the sales call:

• District questions: which district suits a family, what services are around it, and whether it is close to work and schools.
• Price questions: the price per square metre, the yearly rent, and whether the sale is in instalments.
• Size and specification questions: how big the apartment is, how many rooms it has, and whether the villa is a duplex or on one floor.
• Distance questions: how far the nearest metro station is, and how many minutes to the highway or the airport.
• Trust questions: who the developer is, whether their past projects were handed over on time, and whether the project is licensed for off-plan sale.
• Financing questions: whether the project accepts a mortgage, and how much the down payment is.

Here are ten questions written the way customers actually write them, translated from the Arabic they type in their own city's words. They are starting points, not a ready list: swap the districts for your project's and the conditions for what your callers ask.

• Which district in north Riyadh is best for families and close to the metro?
• Apartments for rent in Riyadh near a metro station, 3 rooms: which district fits best?
• What is the price per square metre in Al Malqa right now?
• I want an off-plan duplex villa in north Riyadh with instalments: which projects are available?
• Who is the best property developer in Riyadh for handing over on time?
• Which district in Jeddah is best for buying an apartment near the sea?
• What is the rent for a two-bedroom apartment in Al Rawdah, Jeddah?
• What is the difference between living in Khobar and living in Dammam for a family?
• Are the off-plan projects in Al Arid licensed, and when is handover?
• Which area in Dubai is best for buying an apartment as an investment?

Not necessarily your direct rival, but a page built in the shape of an answer. The sources an engine may cite for a property question come in five kinds, and we name none of them:

• Property listing portals and their apps: many listings, each putting the district, price, size and rooms in one line, which is exactly the shape an engine quotes from.
• Classifieds sites: listings from individuals and offices, short, recent, and carrying an explicit figure.
• Guide and comparison articles: pages like "the best districts in north Riyadh for families" answer the district question head-on, even when nobody in the market wrote them.
• Property news: a story about your project's launch or handover may be cited instead of the project page itself, because it names the date, the developer and the district in one paragraph.
• Other developers and brokers: a rival who wrote one good page about one district may need nothing more to take the whole question.

One page per project, on your own site, answering the four questions at the top. Not a campaign page that dies with the ad, not a designed brochure to download, but text the engine reads and quotes. Here is what the page needs, in order:

• Two lines at the top that answer the question: the project's name, district and city, unit types, price range, sizes and handover date. For example: "Project X in District Y, Riyadh: two- to four-bedroom apartments, from A to B square metres, prices from C, handover in D."
• The unit table as text: type, size in square metres, rooms, price and status (available, reserved or sold). A table on the page, not a picture of one.
• Distances you measured yourself: minutes on foot to the nearest station and its name, minutes by car to the highway and the airport, then the schools, mosques and shopping centres nearby by name. "Close to all services" is quoted by nobody, because it answers nothing.
• The developer and the licence: the developer's name, their past projects and the years they were handed over, and the off-plan sales licence status if the project is not built yet. If you are a broker in Saudi Arabia, write your FAL licence number as text on the page.
• How to pay: cash, a mortgage, or instalments tied to construction stages, with each instalment's share if it is fixed.
• Real FAQs: the questions callers actually ask, each as a heading with a two-line answer under it.
• A last-updated date: written on the page, and changed whenever a price changes or a unit sells.

You ask the engines your customers' own questions, read the answers the way the customer did, and repeat that on a schedule. A question asked in a single chat is not enough, because the answer changes from one time to the next; that is why we measure every question repeatedly, so one lucky answer is never counted as a change. In Tasadr the steps are five:

1. 1Add your project's or your office's site, and set its country and dialect: Saudi Arabia if that is where your customers are.
2. 2Write ten questions of the kind listed above, or pick from the suggested ones.
3. 3Read each question's result on each engine: cited, mentioned without a link, absent or no answer, with the answer's text as the engine wrote it and the links it cited.
4. 4Read who was cited instead of you, and open the competitor analysis that puts your page beside theirs and says what is on theirs and missing from yours.
5. 5Edit one page, then wait for the next measurement cycle, or start an on-demand run after the edit.

Start from the number of sites, then the number of questions. Each site has its own country and dialect, and questions are a balance for the whole account rather than per site. Prices are in Saudi riyal and all published on Tasadr's plans and prices.

• Free: one site and 5 questions on ChatGPT and Google AI Overviews, with no credit card. Enough to see where one project stands today.
• Starter at 499 SAR a month: one site and 30 questions on 3 engines, adding Gemini, with a new analysis every week. It suits a brokerage covering its main districts from its own site.
• Pro at 999 SAR a month: 3 sites and 100 questions on 5 engines, adding Perplexity and Google AI Mode. It suits a developer with a site for each project, or an office with more than one brand.
• Business from 4,999 SAR by request: 10 sites and 150 questions on 8 engines, with a new analysis every day. It suits a developer with projects in more than one city, or an agency marketing several developers.

Six mistakes that keep a property page out of the answer, all fixable in a week without a web developer:

• "Price on request" instead of a price range. A buying question asks for a figure, and whoever writes it takes the question.
• The price and size table and the unit floor plans locked in images or a designed brochure, so the engine finds no text to quote.
• "Close to all services" instead of a measured distance to a named station, road or school.
• A campaign page switched off after the ad ends, so the link that was cited leads to an error page.
• Sold units still listed as available, and last year's prices with no updated date, so the page is read and then dropped.
• District names in English only or in a spelling customers never type, and measuring in English for a market that asks in Arabic.

Nobody can guarantee that ChatGPT will recommend your project. The engine picks its sources afresh on every request, and nobody sells a place inside its answer. What we can do is measure regularly and tell you where you stand today, who took your place, and why their page is clearer than yours.

We do not promise calls or deals: we measure your place inside the answer, and the Tasadr score is ours, not OpenAI's and not Google's. We do not review your listings against property regulations; that is your legal adviser's job, not ours. Other questions about measurement, plans and data are answered in the frequently asked questions.