Engineer Topics:Big Data & Rakuten AI Platform

Engineer Topics
Big Data & Rakuten AI Platform

Create new global value through Big Data x AI

The Rakuten Group has assigned persons to promote the use of artificial intelligence (AI) in all of its businesses in order to meet the needs of constantly changing lifestyles around the world. They work with data scientists who handle big data to pursue innovations using big data and AI, and they are starting to provide new value to customers.

Leading innovation through the world's top big data

Big data is one of Rakuten's greatest strengths. At this very moment, vast amounts of data, including from e-commerce, fintech, advertising and investing, are being gathered at Rakuten. By aggregating, organizing and using this data, we are contributing to improved architecture.

One example is Rakuten Ichiba. Data is analyzed without identifying individuals, such as a user's browsing history, and then personalized according to each attribute. Merchants on Rakuten Ichiba deploy this information at the right time for the right customers, such as in the form of coupons. Meanwhile, big data is used to provide Rakuten Ichiba's customers with optimized category pages, making it possible to access the products they are looking for now. By aggregating big data daily, Rakuten is looking to create services that fit everyone and make them happy.

Rakuten ID and the vast amounts of data on consumer behavior analysis based on it are a source of new service development. Rakuten Ichiba analyzes big data concerning consumer behaviors on its pages using statistics and machine learning. It provides to merchants a page diagnostics service that detects places where customers spent a long time browsing or when they left a page. After optimizing pages, a test was introduced for 1,100 merchants, with each achieving an average conversion rate of 43%.

Big data analysis is brimming with dynamic information for improving technologies, such as demand forecasting and optimization of targeting, not to mention discovering needs of search functions using intuitive keywords such as "kawaii" (cute). Rakuten's engineers, using the point of view of data scientists, will continue to be among the first to create new value.

Making a new major platform a reality using AI

Rakuten established the AI Promotion Department in 2016. Later, in April 2017, it launched the Rakuten AI Platform, aiming to introduce AI into all of the Group's services. A team of around 30 talented engineers from around the world moved ahead with development quickly, and as a result, as of 2018 Rakuten had more than 30 services where it has successfully introduced AI.

Rakuten Ichiba has already launched an automated response service using AI, which answers simple questions, such as "How do I log-in?", 24 hours a day 365 days a year using AI.

The next phase of AI is supporting human decision making. Rakuten is building a system where chatbots will talk with customers just like an actual salesperson and recommend specific products. Ultimately, Rakuten envisions "collaborative AI-nization" where multiple AI work together to satisfy people's needs. Currently, the company is moving ahead with development and reviews to make this a reality.

Simple work will be gradually shifted to AI while creative work only possible by humans will be expanded. Following this development concept, Rakuten will continue to build a new AI business model in ways only it can.

Create next generation businesses
through Digital Signage Award winning technology

Rakuten's AI promotion includes proposals of next-generation business models. In June 2018, Rakuten Institute of Technology had two solutions win an award in the interactive category at the Digital Signage Awards, which celebrates excellent digital signage. Winning the awards was an AI Janken Machine and Virtual Styling Labo. The reason for this double win was the natural combination of digital signage and AI with actual services used in society.

The AI Janken Machine is an interactive signage where a person can play "janken" (rock-paper-scissors) with a character onscreen. Combining a 3D camera, along with trained machine learning and image recognition feature, it can determine a person's move at high speed. It plays janken by first waiting for the opponents move using a speed not possible with the human eye, and wins nearly 100% of the time. The AI Janken Machine was awarded for its entertaining nature and possible applications in business.

AI Janken Machine