April 2018 marked one year since the launch of the “Rakuten AI Platform.” The system, a result of a partnership between Rakuten and IBM Japan, was designed to facilitate the introduction of chatbots into the customer support framework of Rakuten’s many online services.

Since its inception, the platform has taken off in a big way. As of April 2018, the Rakuten Group has welcomed more than 30 chatbots to the family to help customers do everything from calculate shipping fees to look up their favorite baseball players.

A major pillar of customer service

Some chatbots have been developed in multiple languages to cater to Rakuten’s global customer base, such as this example from the Rakuten Global Express service.

Some chatbots have been developed in multiple languages to cater to Rakuten’s global customer base, such as this example from the Rakuten Global Express service.

A large source of demand for automated communication has come from Rakuten’s many customer service departments. With over 90 million unique members in Japan alone, strong customer service is a major area of focus for building a respectable brand and earning customer trust.

The speed of this shift to AI is most evident in Rakuten’s online shopping business, Rakuten Ichiba. Since October 2017, AI Chat has taken care of more than half of all customer inquiries on the platform, drastically improving waiting times and service quality. Other customer service-intensive businesses such as Rakuten Mobile are catching up fast.

“These businesses usually have hundreds, thousands of FAQs, and users need to find the answers from among those,” says Masayuki Chatani, who has led the company’s AI strategy since the beginning of the program. “By leveraging AI, we can provide users with the responses they need quickly, 24 hours a day.”

The response so far has been encouraging. “Use of the chatbots has been rapidly increasing,” Chatani says. “The more these businesses grow, the more inquiries there are – that’s why it’s important to be able to scale like this.”

AI in Rakuten: Only the beginning

In recent months, other services have been steadily adding to Rakuten’s chatbot library with tools for checking Rakuten Points, confirming the delivery status of products ordered online or helping business partners such as hotels. But Chatani sees the potential of AI to reach much further than simple question-and-answer functions.

The official Rakuten Eagles chatbot uses Rakuten Viber to give users rich information on various Eagles topics such as player profiles, results and videos.

The Rakuten Eagles chatbot uses Rakuten Viber to give users rich information on various Eagles topics such as player profiles, results and videos.

“Search is another area where AI can play a serious role,” he explains. “With traditional search, you need to input the appropriate keyword in order to find the product you want. AI can handle more vague questions, such as ‘I want something cute,’ or ‘a gift for my wife.’ It can transform interactions between humans and computers from something computer-centric to something more human-centric.”

While the chatbots are currently restricted to text-based communication, Chatani is pushing for the development of voice-based interaction to make them even more versatile. He sees additional sources of information being integrated – past the basic FAQs to intricate product knowledge and personalized user information. He also sees chatbots expanding past the current web interface onto platforms like Rakuten Viber and other native apps.

Responsible AI development

Chatani is aware of the potential problems associated with entrusting decision-making to artificial intelligence, but emphasizes that the enormous benefits of the technology cannot be ignored. “There needs to be a balance between the benefits and the drawbacks of such new technologies… Look at automobiles, for example. There are drawbacks to allowing everyone speed around in cars, but the benefits are far larger.”

As for the prospect of replacing human jobs, Chatani compares AI’s evolution to that of trains in Japan. “30 years ago, there were people at the stations checking every ticket,” he recalls. “Now that’s handled by machines. The people who were checking tickets are being empowered to do other jobs, like helping passengers, and I think that is a positive development.”