[INTERVIEW] CRE Korea uses AI, big data to offer accurate realty appraisal

CRE Korea CEO Raymond Chetti speaks during an interview with The Korea Times at the company's office in Yongsan-gu, Seoul, Oct. 7. / Korea Times photo by Baek Byung-yeul

By Baek Byung-yeul

An increasing number of Koreans are using smartphone apps, such as Zigbang or Dabang, when they purchase, sell or rent apartments and other residential properties.

But things are different when it comes to dealing with offices and other commercial real estate properties because there aren't many online tools available for sellers and buyers. So they tend to depend more on realtors for market information, forcing them to pay higher commission fees.

In addition, many commercial property owners do not know exactly how much their properties are worth, according to CRE Korea CEO Raymond Chetti, who said the owners have to pay significant amounts of money to measure their value.

"The nation's automated appraisal service for commercial properties was still in its infancy when I was working as a real estate agent," Chetti said in an interview with The Korea Times. "The widely used automated valuation model (AVM) here is not as accurate as similar ones in the United States and other countries."

In a bid to take advantage of the emerging appraisal service market, Chetti said he established his company in 2018 and developed its own AVM service based on artificial intelligence (AI), big data and machine learning technologies.

With its AVM tool, Chetti said he would like to bring transparency to the traditionally opaque real estate appraisal market.

"Before I established CRE Korea, I worked at a local real estate company selling commercial real estate properties such as high-rise buildings. While working as a real estate agent, I found that the commercial real estate market is very challenging for both sellers and buyers because there was a huge price gap between what building owners are seeking and buyers are willing to offer," Chetti said.

"Using AI, big data and machine learning, CRE Korea offers accurate value and predicts commercial real estate trends. With our tool, customers can take a look at the current real estate price, the price after three years and properties with a similar price nearby."

CRE Korea's automated evaluation tool is comprised of three services. The first is a value report which tells the current market price, land information, rent lease status and transactions that have similar demand and supply conditions.

The second is an appraisal service for sellers and buyers as the company reprocesses raw data provided by the Ministry of Land, Infrastructure and Transport and the Korea Appraisal Board, using emerging technologies such as machine learning, AI and data science.

The third one is the data and analytics service, which aims to help financial institutions, banks and investors make an easy and accurate decision on a certain property.

CRE Korea has provided its AI and big data-based valuation tool to real estate service providers. The latest partnership was inked on Oct. 1 with REMAX Korea as they signed a memorandum of understanding to cooperate on reprocessing real estate data provided the local unit of the global estate service franchise.

"Based on real estate data of REMAX Korea, we will provide an appraisal service so that users can get a more accurate valuation," he said.


Chetti said the 11-staff company was able to develop its AVM tool to be well-received by real estate service providers as the company is equipped with an experienced team of data science experts, real estate experts and software programmers.

"Compared with other fintech startups that offer similar AVMs, our tool is much more accurate so that users can have crisp and quick information regarding the commercial real estate market here. The company, which was kicked off with an investment amount of only 150 million won, now has a tool that has an error rate of between 5 percent and 10 percent," he said. "We will try to lower the error rate to 2 percent to 3 percent, which is similar to U.S. residential real estate data analytics company HouseCanary."

The CEO said the AVM tools can help increase the transparency of the real estate market in Korea and stimulate market growth.

"Korea's real estate market is not transparent at all. According to data provided by real estate firm JLL, Seoul placed 73rd in the 2018 global real estate transparency index, followed by Shanghai at 75th, Beijing at 76th and Bangkok at 77th. Its rival cities such as Singapore (48th), Hong Kong (49th) and Tokyo (50th) were all ranked way above the capital city of Korea," he said.

"I believe our AVM tool can contribute to making the country's real estate market more transparent, which can eventually bring it to the attention of an increasing number of investors."

Adopting explainable AI to appraisal service

The CEO said CRE Korea is also working on adopting explainable AI (XAI) technology to its business model to offer upgraded appraisal services to its customers.

XAI is an emerging term in machine learning capable of providing answers that can be understood by humans, while the traditional AI technique is limited by the machine's inability to explain its thoughts and decision-making processes to human experts.

"We will soon change our company name to XAI LAND because we will eventually adopt explainable AI techniques to our real estate appraiser service. In order to emphasize that we are doing both AI and real estate service we decided to go by XAI LAND," he said.

While CRE Korea is currently focusing on offering commercial appraisal services, Chetti added the company also plans to launch a valuation service for residential real estate under the new company name.

"Whenever I met with real estate experts from financial institutions, many of those experts thought highly of our service and always asked me whether CRE Korea has a similar service for residential real estate and I answered yes we will," he said. "We will launch a non-commercial appraisal service under the new company name XAI LAND."


Baek Byung-yeul baekby@koreatimes.co.kr

Top 10 Stories

LETTER

Sign up for eNewsletter