Alex Chen Ye stands out, as usual, in a sea of sober-suited professionals and entrepreneurs at an artificial intelligence (AI) forum in his unpretentious T-shirt and sweatpants combination.
It’s his personal style whenever he appears in public, explains a colleague of his.
Likewise, Chen also keeps a low profile but being strong-minded in the workplace. The former senior vice-president of advertisement at Beijing-based Meituan-Dianping last year left the on-demand online service giant, which just listed in Hong Kong, and jumped onto the financial technology bandwagon. In mid-2017, just months after he quit, Chen founded Tigerobo, an AI-based financial information search engine headquartered in Shanghai.
AI technology has made big strides like never before and boomed last year, with Chinese mainland internet giants like Baidu, Alibaba and Tencent launching their respective strategies in the field.
Chen and his partner John Canny — a computer science professor at the University of California, Berkeley — were excited to see the advance of AI in China, partly driven by the development of deep learning — a set of techniques that are able to predict accurately after being married with large amounts of data.
“Deep learning enables many AI tasks to reach their inflection point and be usable,” says Chen.
Having been committed to the AI field for years, Chen and Canny pictured quite a few AI applications and finally decided to move forward in the financial information sector.
The marriage of AI and finance usually leads to three directions — robo-adviser, smart customer service, and intelligent search. Among them, robo-adviser has gained most favor, known as the AlphaGo in the investment world. And smart customer service is developed vigorously by major corporations such as Ping An Insurance Group.
However, Chen’s eyes were on intelligent search, seeing it as an underlying technology.
Intelligent search is commonly based on one element — text language. With AI, it has already been demonstrated that images can help on security, while language can be applied to user interaction. Afterwards, Chen deemed it’s the right time for text language boost, such as text processing and natural language processing (NLP).
NLP is a field that AI teaches computers to understand human written language. It’s being widely developed and used by search engine companies like Google and Microsoft to comprehend their users.
“The market for text processing and NLP will be many times bigger once they find a business model because the majority of human knowledge exists in terms of text,” Chen reckons.
Tigerobo aims to be the “Google” in the financial information search field. Compared to traditional financial information provider Bloomberg and Reuters, Tigerobo strives to offer simpler and more direct experience.
With the company’s products, users only need to ask with a single complete sentence instead of inputting commands or key words.
Chen cites one of the functions — stock selection, which is part of robo-advisor’s job. If you ask: “Which stock performed well in the past three years with a low P/E ratio and robust earnings?,” the search engine will help you pick some eligible stocks in one second.
“We chose to start with something related to text processing and language comprehension, as they are so fundamental that can be applied to other areas easily,” says Chen. “If we would like to try something else in future, it’ll make the transformation smoother.”
Tigerobo has so far released four products for its clients, covering banks, funds and securities traders, to meet customers’ different needs.
The services provided include corporate information, industrial research, market analysis, public opinion monitoring and some predictions. Internal enterprise search is also accessible for those data-based companies by organizing and structuring huge amounts of data that the clients bought, but found no way to integrate.
Meanwhile, the only product currently available for retail clients is a WeChat application named “prospectus”, which monitors Chinese companies that are about to go public in all markets.
“The system can detect a company’s prospectus just one minute after it’s submitted online, fetch figures and key points from hundreds of pages, and then generate a piece of short news automatically,” Chen explains.
He says they’re also working on a product translating long-form financial stories by foreign media into short pieces in Chinese. Sources and original links will be provided.
The startup completed its Pre-A round financing of more than 100 million yuan ($16 million), led by NWS Holdings and Gaorong Capital, in June this year. Sun Hun Kai & Company also took part in the fundraising.
Chen says the new funds will be used in three fronts — manpower, database purchase, as well as marketing and promotion.
Hong Kong focus
Besides an existing research and development center in Shanghai, Tigerobo is scheduled to set up another two R&D centers — in Beijing and Silicon Valley in the United States.
At the same time, considering almost half of Tigerobo’s business clients are based in Hong Kong, the company aims to build up a team in the city to take charge of services before and after sales.
“Business clients in Hong Kong are more willing to pay for the services as the demand in the city is stronger. We help them on asset digitalization and consumer engagement, so it’s easy to gain from this part,” says Chen.
Just the same as a traditional financial information terminal, users need to pay a yearly subscription fee to enjoy the service, but at a much lower price, plus a deployment fee in the beginning.
After attracting a school of loyal clients, Tigerobo could also cooperate with certain related platforms, convert online traffic to those platforms and monetize from it, Chen adds.
The entrepreneur is eyeing a Series A funding round as soon as within half a year. As for a listing, he admits it has to wait to see the progress of its products, as well as market opportunities, but he prefers foreign capital markets.
“The market is so fresh that our focus will be on how to attract more users at present.”
Contact the writer at [email protected]