How e-commerce giants are using AI and marketing (Part 2)
Columnist Daniel Faggella explores how e-commerce leaders like eBay, Rakuten, Flipkart and ASOS are utilizing AI to power improved product search and automated customer service.
The explosion of AI in e-commerce has allowed online retail giants to explore advanced technologies. As firms in the e-commerce industry continue to expand and cater to a growing number of customers, the need for an automated system to streamline their operations and marketing has become one of their top investment priorities.
In this article, Part 2 of a series, we’ll look at how eBay, Rakuten, Flipkart and ASOS are employing artificial intelligence. In Part 1, we discussed AI applications at Amazon and Alibaba.
Similar to Amazon and Alibaba, eBay has also invested in predictive model research to reduce the time its users spend finding preferred products. To further compete with Amazon, the California-based e-commerce firm acquired learning engine startups SalesPredict and Expertmaker in 2016.
The company believes it will strengthen its structured data through the acquisitions’ predictive analytics and machine learning capabilities. Amit Menipaz, vice president of structured data, characterized the benefits of these additions like this: “For our buyers, it will help us better understand the price-differentiating attributes of our products, and, for our sellers, it will help us build out the predictive models that can define the probability of selling a given product, at a given price over time.”
eBay also continues to invest in machine learning and data science through eBay Research and sponsorship of research events and conferences. Earlier this year, the company introduced Image Search — an app similar to Amazon Rekognition — which matches a user’s photo of any item to visually comparable listings from the 1.1 billion items featured on the website. If the user prefers to use any photo on their social media account or from a web browser, they can use it to search for similar items using Find It On eBay instead, which was introduced at the same time. These two apps were developed using computer vision and deep-learning technology to generate quick results and improved product matching.
In addition, interior designers or amateur home decorators can use eBay Collective to shop for high-end home items featured on the website. The Shop the Room feature lets users hover over items shown on each room image, and eBay will display similar items on a pop-up window — all powered by AI.
And going beyond keywords to search for products, eBay launched eBay ShopBot for its users. The app runs on Facebook Messenger and lets shoppers interact with a chatbot using text message, speech or image. According to eBay, the goal of the app is to provide the services of a personal shopper by giving tailored search results.
Japan’s largest e-commerce site, Rakuten, continues to look into ways it can better predict buyer interests through AI. With the Rakuten Institute of Technology, launched back in 2006, they are able to use data from its 200 million products to forecast product sales with a high degree of accuracy, the company claims. They also say they can segment buyers more accurately using real-time data. Image-recognition technology through AI is applied in the company’s fashion recommendation app, Rakuten Fits Me.
Sunil Gopinath, CEO of Rakuten India, explains in an interview with Factor Daily that they use the same technology for different applications. Additionally, he said, “Out of over 70 services we have, almost 30 services will be enabled with AI chatbots by the end of . The objectives are to improve customer satisfaction significantly and also sales productivity. We’re partnering with IBM Watson to enable those AI chatbot engines.”
It also partnered with Sentient Ascend to develop massive website and landing page design-testing tools that are said to be much faster than traditional A/B testing. Ryugen Shimizu, managing director of Rakuten Marketing Japan, said in a statement that he believes that their collaboration is important in order to “accelerate our business, help our advertisers significantly increase their conversion rates and make it possible to implement Plan-Do-Check-Act cycle in marketing activities at speeds that are not humanly possible.”
In the future, Rakuten is looking into finding more uses of its AI technology such as the AI machine, Cassis, which responds to human movement. In its initial testing, the app was able to play rock-paper-scissors through its Kinect 3D sensor and deep-learning technology. Its processing time to determine what hand gesture to show in order to win takes less than a second and makes it unbeatable, according to the company.
India’s homegrown e-commerce company, Flipkart, has an answer to their competitors’ AI developments. By feeding local data to their AI, the company believes that it can better serve the needs of Indians by asking them relevant questions through their app.
The company’s shopping app, developed under what Flipkart calls Project Mira, is still in progress since its February 2016 launch, but it has been promising so far, according to Ram Papatla, vice president of product at Flipkart. In an interview with the Times of India, he said, “What’s interesting is [that] we cover 50 percent of search volume through Mira, and through Mira, we are currently seeing 12 percent cart additions. We are seeing a dramatic improvement in seller experience as well.”
To improve its AI developments, the company has started working with Microsoft and exclusively using its cloud-computing system, Azure. Microsoft made a significant investment in Flipkart so it can cooperate on developing voice recognition and better customer targeting.
Flipkart also is employing artificial intelligence in its online fashion business, built through the acquisitions of Myntra and Jabong, which allowed it to control 70 percent of the online fashion market in the country. Its fashion-related AI projects include chat support services and customized shopping webpages.
Working on artificial intelligence has brought improvements in customer service, according to the company. A report in the Hindu Business Line on the company’s net promoter score (a measurement of a customer’s likelihood of recommending a company’s services or products) claims an increase of 14 percent in the middle of 2017, growth of 14 percent in customer resolution and a drop of 25 percent in customer pain. The company attributes these improvements to chatbots that were launched in April 2017.
Fashion retailer ASOS, one of the UK’s online retail giants, is not threatened by the growth of Amazon, according to a Telegraph interview with its CEO, Nick Beighton. Currently, the company is continuing its developments in AI and voice recognition systems to influence buyer behavior. It has also forayed into image recognition technology by introducing a visual search capability which allows the app to match users’ photos with clothing sold online.
But its technology still has issues. Before leaving the company in 2015, Pete Marsden, the former CIO, told ComputerWeekly that he was working on demand forecasting tools, search capability and expansion in China. He says that handling big data has its challenges. “So understanding what the true data sources are, tidying it up, consolidating data warehouses and so on can be very difficult. It can also be expensive … so cleaning up data can be costly and time-consuming,” he explained.
Though image recognition, in general, isn’t especially user-friendly, ASOS execs believe it can be salvaged by big data and machine learning in the coming years. Still, their technology has a long way to go, which is why the company is reportedly adding to the team of 900 that already work on its mobile app. Coming next, according to Glossy.com, is a mobile virtual assistant called AVA, which will help customers search while serving as a data repository for user preferences and data.
Concluding thoughts on artificial intelligence in e-commerce
E-commerce giants are continuing to invest in AI to find solutions to their product pains, to trim down operational costs, to improve the shopping experience through more accurate product matching and to generate authentic product reviews. AI tools such as image and voice recognition, chatbots and cloud-based assistants are designed with improved machine learning, computer vision and natural language processing capabilities.
However, the technology is still far from being perfect. E-commerce firms continue to improve their AI tools to better match market demand. Companies also collaborate with other organizations to merge their competencies in AI and create more sophisticated tools.
We will continue to monitor how the e-commerce sector evolves over time, as the field continues to be impacted by AI.
Opinions expressed in this article are those of the guest author and not necessarily MarTech Today. Staff authors are listed here.