Dada Now General Manager Bing Fu outlined the Dada Group company's growing efficiency and fulfillment for customers, merchants and riders in a speech entitled "AI empowers on-demand delivery infrastructure" at the three-day World Artificial Intelligence Conference in Shanghai which concluded on Saturday.

General Manager of Dada Now Bing FU shares insights at WAIC.

Dynamic matching

AI technology has gradually matured, Fu said. Online and offline channels are integrating with on-demand retail development. Demands and scenarios for on-demand retail have multiplied alongside a rapid growth in orders.

On-demand delivery orders grew explosively from 2014 to 2021, according to the 2022 On-demand Delivery Service White Paper released by Log Research and Dada Now. There were 30.85 billion orders in 2021, with a compound annual growth rate close to 60 percent, he said.

While more traditional logistics involve days, local on-demand orders are “omni-channel, full-scenario, full-category and full-time, with a requirement to be delivered within one hour,” Fu said.

Delivery capacity has had to be constantly updated and improved to meet the increasing number and variety of delivery orders, he noted.

On the one hand, more riders are joining and their working time and locations are constantly changing, posing a challenge for high-quality management and coordination, he said. On the other, on-demand delivery has evolved into a diversity of types: two-wheel, four-wheel and autonomous delivery.

The key question for the on-demand delivery industry was how best to achieve efficient, stable fulfillment.

"The supply and demand sides of on-demand delivery are more random than traditional logistics and AI technology is essential to match the two sides dynamically,” Fu said. “Meanwhile we also care about the feelings of customers, merchants and riders, while improving efficiency.”

Last mile of delivery

Dada Now's smart order dispatch system can dynamically match a massive number of orders and riders in real-time: 176 million routes a day, with an average order response time within 10 milliseconds and average delivery time about 30 minutes, he said.

The AI algorithm also factors in rider experience and safety. Riders may encounter a diverse variety of circumstances during the process of picking up an order, delivery and handing the order to customers.

AI identifies potential gremlins and delays en route, dynamically extending the delivery time for riders.

Working with JD Logistics and vehicle developer White Rhino, Dada Now's autonomous delivery operation system has fulfilled nearly 50,000 Seven Fresh, Yonghui Supermarket and other stores’ orders in a dozen neighborhoods, Fu said.

Unmanned vehicles mean lower delivery costs, larger carrying capacity, stronger stability and higher safety for one-hour delivery, he said.

"From the integration of AI and on-demand delivery, we can see the AI is making delivery smarter and offering a warmer experience,” Fu said. “With technological innovation, we guarantee that our customers receive their orders within one hour and ensure the riders' efficiency, experience and safety.”