Benoit Galley, founder of Visit, wonders why the use of artificial intelligence in real estate is hampered.
Artificial intelligence, everyone talks about it, many claim to use it but in fact, rarely use it intelligently … In the world of “machine learning”, to create a useful service for users, you need to realize a few marketing The key idea is far from the terminology of speech or data scientists.
Conventional computing or artificial intelligence?
Let’s get rid of the philosophical answer and put the user back in the center. Take the example of a self-driving car. Not just the road to adapting to the code, but the computer code that must be understood and adapted to the outside world. On the contrary, In developing a website, the user must understand the designer’s thoughts, click where he is expected to click and enter his data (title, name, address, etc.) in the dedicated area. So we can say that with artificial intelligence, it is a code that must adapt to the outside world, whereas in classical computing, it must adapt the outside world to the code.
As artificial intelligence is forced to adapt to the outside world, it is condemned to conduct an indicative multiplication in special cases related to the diversity of the living world. Tesla, during his conversation AI for complete self-driving The perfect example of autonomous driving advance. A goal Descending Identifying a stop sign has become as much a test of reality as human innovation.
AI still needs to develop real estate business
The real estate business has changed a lot in recent years. A number of tasks have been automated to save time for both the real estate agent and the user. As such, the profession has equipped itself with a platform to store property information for sale but also profiles of people looking for a property, their criteria, their budget, their situation and so on. In light of this new information gathered, sites have been reviewed to make it easier to set up a property for sale and to make real estate search more intuitive.
3 uses of artificial intelligence in the real estate industry
Nevertheless, although this data storage can be done on a large scale, it is still classic computing. The emergence of artificial intelligence in the real estate sector is currently limited to 3 distinct uses.
First, r Document recognition Avoids many human errors and saves a lot of time by recognizing and digitizing the information contained in a document by an artificial intelligence.
ThenPrice estimates Allows agencies as well as sellers to quickly get an idea of the sale price of the property. An impressive crowd of past public sales data cross-referencing criteria with current trends and prices makes it possible to model the price of a property in a fairly subtle way.
Finally,Data aggregations, which offers new possibilities for identifying certain information on the Internet, copying and indexing it so that it is accessible to as many people as possible. So we see the emergence of neo-real estate portals that make all real estate ads visible, one of the main frustrations of users looking for a property. It also makes it possible to remove “off-market” ads that are not on any portal to date
Obstacle course for skilled artificial intelligence
If the use of artificial intelligence is not widespread, it requires an upstream education foundation to be error-free. Fortunately, real estate is a sector where this error rate can be temporarily taken into account in the case of artificial intelligence, but it is not the same for others as in the case of autonomous vehicles where the error is not tolerable.
Although document recognition error rates are very low because they are quite standard, this is not the case for price estimates or data aggregation. For example, to determine a good price for an algorithm, it can rely on a lot of data. But if this operation determines 90% of the price of a product like Tesla, there are some exceptions that AI must learn (for example, well-related specifications). Even if artificial intelligence facilitates certain tasks, real estate agencies and their agents must rely on their skills to improve it.
Benoît Galy is a polytechnician, data scientist and founder of Vizzit, a startup that makes all real estate advertising accessible to users looking for housing.