Machine learning (ML) and artificial intelligence (AI) are becoming what smartphones were 15 years ago—the flagship of the new reality, says Cian CEO Dmitry Grigoriev in a blog about the role of AI in the development of ‘smart’ classifieds and marketplaces.

“You need to integrate AI so as not to find yourself in the role of catching up. Millions of people heard about it, thousands played with neural networks, drawing exciting pictures, but only a few found application for the technology and took advantage of its benefits.  Companies, especially those claiming to be technology companies, cannot afford to spend too long pondering over the innovation.”

The implementation of AI/ML technologies has turned into a kind of competition: whoever is the first to understand exactly how this will help speed up or optimize work, reduce costs and improve user experience will be in the lead. For companies engaged in online trading and providing services, such as classifieds, marketplaces or banks, the introduction of artificial intelligence is already a top priority.

Artificial intelligence has been called the “universal equalizer”, says Grigoriev. Judging by research, it boosts KPIs of the high-performing staff by 15%, and by 30% of the weaker ones.  The same applies to companies. Large and successful market leaders and start-ups now seem to be equal in capabilities, which means that competition has instantly intensified, and no one can simply rest on their laurels.

AI as a writer

Cian CEO asks: what specifically can AI/ML offer classifieds to significantly improve their work? One example is filling out sale cards. AI can suggest best options for the title and description. For people putting their apartment up for sale, creating a clear description is not always easy.  Artificial intelligence can help with this or even take over this work entirely. In the future, it will probably be enough for Cian users to enter the address of their property and upload photos; the AI ​​will do the rest, highlighting the features important to the buyer.

Realtors can also delegate to AI writing ads, and much more. The lion’s share of a realtor’s time is taken by calls, arrangements for viewings, and planning meetings. An AI-based assistant can do all this instead: answer typical questions about an apartment, find a slot in the calendar to organize a viewing, and notify the property owner and potential buyer. Preparing a contract, sending queries to the bank or filing paperwork for registering a transaction – all this can be delegated to AI.

AI as a censor

The use of predictive models that consider account history and user behavior patterns allows to identify fraudsters much faster and more efficiently than can be done manually. Cian have learned to use AI to automatically identify 80% of dubious ads before they are reported. Grigoriev says Cian is 30% better than other classifieds in catching such ads, and twice as good in Moscow and St Petersburg. Further automation of ads moderation, he says, will reduce Cian’s operating costs by millions of rubles p.a.

AI search engine

For many years, businesses have strived to use knowledge about their customers to create the most personalized offer possible. The next stage of Cian’s development, says Grigoriev, is the maximum individualization of interaction with the client, based on previous user experiences, including search queries, browsing history, chats, etc.


Grigoriev reminds that Cian was the first among Russia-based real estate classifieds to launch a smart assistant, CianGPT that can independently search for real estate of all types using specified text parameters, as well as answer simple questions that are usually handled by the support service. Cian is aware of the developments in foreign classifieds and moves in the same direction, says Grigoriev.

A future search system will take into account not only the request itself, it will “remember” whether the person has a family, what shops they were looking for, whether they drive a car or like going to the movies. Based on this, an offer will be as much tailored to the buyer’s needs as possible.

The “smart” classifieds of the future will have to second-guess what a user wants, and not just react to the request in the search bar, says Grigoriev. Nowadays, a search query returns thousands of results, and the user simply does not have enough time to view them all. People can spend months looking for an apartment, and no one likes this state of affairs, Grigoriev says. In the future, AI should relieve users of the need to choose from a massive array of properties, 99% of which are obviously not suitable for them. Therefore, searching for real estate will never be the same, Cian CEO sums up.

Print Friendly, PDF & Email

Related Articles