Alibaba Group says it has tweaked its chatbot in preparation for the e-commerce platform's annual shopping festival, beefing up the automated tool's predictive and interactive capabilities.
The Chinese e-commerce operator said the Alime Shop Assistant now would offer 24 by 7 automated customer support and feature better scalability for peak demand. The chatbot's predictive analytics also had been boosted to preempt what consumers might ask and prompt them in their shopping decisions, such as reminding them about discount vouchers.
The tweaks were made in preparation for Alibaba's annual shopping festival on November 11, which last year saw the chatbot handle 100 million customer conversations. These had fuelled 2017 sales that accounted for 15 percent of the day's gross merchandise volume (GMV), which totalled US$25.3 billion (168.2 billion yuan).
Alime Shop Assistant was introduced in March 2017 to support merchants on Alibaba's Taobao and Tmall platforms.
According to the Chinese vendor, some 600,000 merchants on the marketplaces currently tapped the chatbot, which it said could help the retailers slash up to half of their previous call-centre costs. The automated tool currently was offered free to all merchants on its platforms.
L'Oreal Paris, for instance, used the chatbot during Alibaba's June 6 promotional campaign to handle 86 percent of customer inquiries on its Tmall online store. This was a higher capacity to the 49 percent it managed in the previous month. The beauty brand added that the chatbot was able to resolve 63 percent of all customer enquiries.
Alibaba's chief customer officer Wu Minzhi said: "We hope that through intelligence-driven, human-computer interaction, we can enable the merchants on our platforms and help their businesses run more smoothly."
The company added that the chatbot was integrated with natural-language processing and deep-learning technologies, enabling it to make personalised shopping recommendations, share promotional and sales information, revise order details, and manage returns and refunds.
Merchants also could feed Alime Shop Assistant information about their brand and products as well as sales campaigns, to further train the chatbot and customise the tool for their own online stores. Customer queries that could not be resolved would be automatically rerouted to a human assistant.
Alibaba's deep neural network model in January outscored humans in global reading test, Stanford Question Answering Dataset. In July, the Chinese company also introduced an AI tool that it said could produce up to 20,000 lines of content a second, helping merchants automatically generate product information on its e-commerce sites.