知Yizhi Daily
← Back to this edition

Business & markets

Six Embodied AI Robots Begin 24-Hour Trial Shifts at Shanghai Convenience Store

Edition 2026-11-04 · 08:00 Beijing time

A convenience store in Shanghai's Xuhui District is testing embodied intelligent robots in a real business scenario, with six robots dividing tasks over 24 hours, including greeting customers, picking and packing goods, restocking, cleaning, making coffee, and heating prepared food. Researchers said the open scenario can test the robots' and algorithms' capabilities and help explore applications such as overnight duty.

From 10 a.m. on November 2 to 10 a.m. on November 3, six embodied intelligent robots took on divided “trial shifts” at a convenience store in Xuhui District, Shanghai. The robots performed tasks such as greeting and receiving customers, picking and packing goods, patrolling and restocking, cleaning up trash, making coffee, and heating prepared food, undergoing assessment in a real operating scenario.

Ren Guanghui, vice president of the AgiBot Embodied Intelligence Research Center, said that compared with factories, where environments and tasks are relatively fixed, convenience stores are more open, with many types of products and different customer needs, and robots must also use multiple tools to complete delicate operations. Taking packing a steamed bun as an example, gripping the bun and opening the paper bag involve multiple steps such as grasping and tool use.

Opening a freezer is also a research and development challenge. The robot needs to judge the direction in which the door opens and the amount of force to use, hold the door with one robotic arm, and then pick up goods with the other hand; the two arms must coordinate to complete continuous movements. In addition to perception sensors at positions such as the head and wrists, force sensors on the robotic arms help sense the amount of force applied, avoiding violent collisions or damage to the door.

The test covered the noon rush, evening rush, late-night duty, and morning rush, with assessment content including continuous on-duty performance, business capability, performance during each time period, and customer reviews. Ren Guanghui said complex tasks in open scenarios can test the capabilities of robots and algorithms; meanwhile, 24-hour convenience stores have practical needs such as overnight duty, and the team hopes to explore having robots take on jobs that require day-night reversal or are unsuitable for people to perform for long periods, while maintaining their working state through methods such as autonomous battery swapping.

Original sources

中新网财经 ↗
Back to this editionOpen interactive reader