The world’s fastest supercomputers reside in China, not America. And, while the majority of the planet’s most valuable tech brands still carry Uncle Sam’s imprint, Chinese rivals tread not far behind. Alibaba, Tencent, Baidu, Xiaomi, and Huawei are just a few of the massive brands across the Pacific that rival the unicorns in Silicon Valley.
In the emerging and culture-shifting field of artificial intelligence, we seem fixated on developments that occur only in our backyard. In fact, while US media continually lauds the achievements of IBM’s Watson and Google’s AlphaGo, we hardly ever hear of Xiaoi, a conversational AI giant whose bots — deployed to almost every business sector in China — have engaged 500 million users and processed more than 100 billion conversations. Xiaoi bots take on different roles in finance, automotive, telecommunications, e-commerce, and other industries, serving many Fortune 100 companies.
Compared to the fledgling state of the chatbot industry in the U.S., chatbots already serve as a business staple in China, with Shanghai-based Xiaoi leading the pack. Xiaoi was founded in 2001 and serves hundreds of medium to large enterprises, including China Mobile, China Construction Bank, Samsung, Ford, Lenovo, Volkswagen and Huawei.
Xiaoi channels a significant portion of its R&D resources on natural language processing (NLP) and natural language understanding (NLU) which deal with machine reading comprehension. The core Xiaoi AI framework drives the company’s product lines — from customer service bots and interactive voice response (IVR) to IoT solutions for smart furniture, cars, robotic toys and other products.
Headed by Pinpin Zhu, the company pioneered the commercialization of chatbot technologies in China, reportedly controlling 90% of the conversational bot market in the country. With its commanding presence, Xiaoi is also a major force behind the global bot economy. In 2016, PinPin Zhu delivered a keynote speech during the 25th International Joint Conference on Artificial Intelligence (IJCAI) in New York, revealing more about his company’s approaches in technology design and monetization.
Road to market dominance
The first Xiaoi chatbot debuted in 2004 over MSN’s messaging platform. Prior to this milestone, only human beings were able to chat on the channel, according to a China Today interview with Pinpin Zhu. As the chatbot’s pool of users accumulated to more than 20 million, Pinpin Zhu thought of ways to monetize the interactions.
Customer service was the primary enterprise area Zhu hoped to automate using chatbots. In 2006, Xiaoi added specialized field knowledge to the chatbot’s function list, on top of weather reports and stock data. The move bolstered the platform’s capabilities as a virtual customer aide.
Finding customers was difficult at first for Xiaoi, however. The AI-driven solution was too new to gain market traction and most prospects opted to stay on the sidelines. Even then, Zhu was able to make the case with Jiangsu Mobile, Xiaoi’s first corporate client. Initially, the chatbot assisted customers on the telecommunication company’s website, then addressed queries over text messaging. The results were modest but positive: Jiangsu Mobile reduced labor costs by around $300,000.
The road to commercialization became a steady climb since then.
Among the top 50 banks in China, 40 are Xiaoi’s clients, and almost all customer service providers use Xiaoi’s services. In industries where large-scale use of smart customer service solutions is uncommon, the minority users are almost always Xiaoi’s clients as well. Xiaoi processes around 200 million interactions per day, with interactions from the financial sector (banks, securities, insurance) accounting for the largest proportion.
Xiaoi controls 90% of the market. The remaining 10% are serviced by Udesk, Sobot, AiKF, Sinovoice, Huilian Infotech, and other bot technology players.
A case study to bank on
China Merchants Bank (CMB) was the first enterprise to provide real-time chatbot interaction on WeChat. Users can search for information, make transactions on WeChat, or bind their ID with the banking system. They can also perform repayment, transfer money, or redeem their reward points. Meanwhile, the bank can quickly perform user data collection, marketing analysis, promotion activities, and other tasks. It can also use LBS, mobile payment and other functions to provide various services to create more revenue. By the end of 2015, there were 15 million registered users on WeChat. Interactions per month on three channels (WeChat, QQ, APP) exceed 41 million. The chatbot handles 99.5% of the questions with an accuracy of 99.8%.
Xiaoi also used its AI platform to build an NLP-capable knowledge base system for CMB. The system provides employees with a flexible knowledge management solution, making it easier to maintain and distribute knowledge. The platform helps drive internal staff training. On the WeChat platform of CMB University, students can use AI technologies to perform registration, textbook selection, class sign-on, satisfaction survey, and other actions. Students can also book flight tickets or hotel rooms when they need to train elsewhere.
Making chatbots work at home
Commercializing conversational technologies remains a huge challenge in the US, with many enterprises still in doubt of either their company’s AI readiness or overall chatbot effectiveness, security and ROI. A few players such as Nuance, IPSoft, IBM Watson and Kasisto have made strong business cases for chatbots but they hardly scratch the surface when compared to Xiaoi’s traction.
At most, leading conversational AI providers in the US have been in the bot market for three years. Perhaps it’s time for you to have a chat with Xiaoi, who has been in the chatbot game for over fifteen.