Most property sites ask you to fill in a form before they show you anything: a price band, a number of bedrooms, an area. Then you scroll through the results and sort them in your head. A plain-English search starts from the other end: you write the question, and the tool works out what to look up.
From forms to full sentences
A form works well when you know what you want in the form’s terms. It works less well for a question such as “which towers in this area sold the most flats this year?” A listing page shows what is for sale, and that question is about what has sold.
Plain-English search takes the question the way you would put it to a colleague. Some property data platforms now offer it. Ask AI on dataHabibi, a Dubai housing data platform, is one example: it takes property searches written in everyday English.
The difference shows in what comes back. Scrolling gives you a sample of the market, in an order the site chose. A direct question gives you an answer drawn from all the records the tool holds: how many, how much, how often. It also exposes your assumptions sooner. If a sharp question brings a surprising answer, you learn something that a list of results rarely shows.
The change is small but real. You spend less time translating your question into filters and more time reading the answer.
Questions worth asking
A good question names three things: the place, the type of home and the period. Here are five that a property data tool should be able to take:
· What have two-bedroom flats in this tower sold for over the last twelve months?
· Which towers in this area recorded the most sales this year?
· How does this tower’s price today compare with its launch price?
· Which areas show rising prices over the past two years?
· What rental yield have one-bedroom flats in this building produced?
Compare these with “Is this a good area?” The vague question has no answer you can check. The specific ones do.
Say what shape of answer you want, too. Ask for a count, a range or a top three, and not only an average. A count tells you how far to trust the rest, and a range shows how much the prices vary. If the tool allows follow-up questions, “Show me the sales behind that” is often the most useful one.
Check every answer against the records
Treat the first answer as a lead, not a verdict. Three checks take a few minutes.
1. Read your question back. Did the tool use the area, the flat type and the period you meant?
2. Ask what the answer rests on. A sound answer points to sales you can look at, with dates, sizes and prices. Check two or three by hand.
3. Look at the count. An answer drawn from three sales is a hint, not a pattern.
Take an invented case. You ask what one-bedroom flats in a tower sold for last year, and the tool replies with about AED 900,000 from 14 sales. The figures are made up, and the checks are simple. Are the sales one-bedroom flats? Do the dates fall in last year? Are there 14, or only two? If the records match the answer, you can use it. If they do not, you have found a problem before it cost you anything.
If a tool cannot show what an answer rests on, do not rely on that answer.
Where plain-English search falls short
Words are loose. “Cheap”, “quiet” and “close to the beach” mean different things to different people, and a tool may pick a meaning you did not intend. A tool also knows only what is in its data. It cannot know how a flat is finished, what the neighbours are like or what a seller will accept.
A smooth sentence can also make a weak answer sound firm. Keep reading the numbers and not the tone. When a question is vague, narrow it: name the building, name the flat type, name the dates.
Plain-English search saves time at the start of research. It does not replace the checking at the end. To practise the three checks on a real building, try a question in Ask AI at datahabibi.ae.
