How AI is transforming JD’s retail business

During China's '618' online shopping festival last June, online retailer JD.com saw the use of artificial intelligence technology drive revenue growth by more than seven times year-on-year and contribute to a 145% hike in conversions, according Bowen Zhou, president of JD AI at JD.com. 

Speaking at Rise conference in Hong Kong yesterday, Zhou said AI technologies can help retailers in three fundamental ways: cost saving through automation; using AI to augment business decisions, operations and management; and most importantly, drive new market entry by satisfying consumers' increasing demands for a continuous supply of personalised interactions and content.

High efficiency

JD has deployed AI in almost every aspect of its retail process, Zhou said, from smart consumption to smart manufacturing, and smart distribution to link them together. The increasing interaction between the consumption and manufacturing units has led to supply chain innovation. To empower these different components, JD created an open AI hub platform, providing a basic layer of AI capabilities and solutions to support a wide range of functions.

One such function is helping drive logistics innovation, Zhou said. At the last 618 festival, JD rolled out an AI bot to make all outgoing calls to customers ordered large appliances such as big-screen TVs. The bot asked consumers to book delivery and installation appointments, and Zhou said the interaction was so natural that many people didn’t recognise it was AI.

Moreover, the bot system helps detect background noises and consumers' emotions during the conversation, so that when they are not in the mood to discuss delivery or other issues, the machine automatically hangs up and asks them to call back at another time.

Zhou said the AI chatbot service has helped improve efficiency significantly, giving JD's staff more time to solve more complicated problems for customers.

Better user experience

JD is also enhancing the entire shopping experience with a personalised AI assistant, which can be supported by human staff to ensure a good experience. For example, the AI assistant can interact with customers to help them make decisions as to what to buy.

Zhou said the assistant is expected to produce good answers to customers’ questions, as it can understand their needs and intent by having conversations with them and consuming lots of customer feedback information from JD’s website and other sources.

“We see very positive signs using AI as a shopping guide to improve conversion rates and help brands improve [their] revenue. Also, for the post-sale process, our AI chat bots are handling millions of customers every single day,” he said.

AI-powered checkout system

New scenarios

On top of these examples, Zhou said AI can also drive new market growth by helping creative innovative new retail solutions.

One recent example is JD’s partnership with Fung Retailing Group to create an AI-powered checkout system in Hong Kong, which allows consumers to grab multiple items and put them all together on a machine for quick checkout. Zhou said it is expected to reduce waiting times by 30% during rush hours.

While JD’s competitor Tmall, owned by Alibaba Group, recently announced upgrades to the flagship stores on its platform that involve more video and livestreaming choices for shoppers, JD is also enabling customers to shop using different touchpoints such as images and videos, rather than simply search.

But for all the visible benefits brought about by AI, it is the data powering it that is more important for marketing, Zhou said, as it provides greater guidance for the brand. Thanks to its technological improvements over the past year, JD can gain even more information about its potential consumers. Zhou said JD's data shows that majority of the users of its AI-related services are born after 1995, a significant portion of them come from China's lower-tier cities, and they are buying fashion and beauty products.

"We hope to collect all the intelligence we learn using AI and feed it back upstream into manufacturing, design, marketing, and even to local governments that can use this intelligence to optimise local services," Zhou said.