Learning Robots During Business Hours: 24-Hour Convenience Store Demonstration by China's AGIBOT
Learning Robots During Business Hours: 24-Hour Convenience Store Demonstration by China's AGIBOT
Chinese robotics company AGIBOT demonstrated its 'Genie Master' AI foundation and AgiBot G2 humanoid robots in a Shanghai convenience store from November 2 to 3, 2026. The six robots performed tasks including restocking, serving coffee, and cleaning, with the company reporting 1,730 total tasks and four robots passing evaluation criteria. The demonstration highlighted the system's ability to adapt to on-site failures, such as switching camera views when shelf tags obstructed a robot's view. Independent verification of human intervention levels, maintenance time, and exact success metrics is not available from public sources.
Source: Robotics — Robot learning news discovery · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
tasks
1,730
Reported trials
View original evidence
in Shanghai, China. They tested multiple tasks that occur in a store, such as restocking, handing over products, serving coffee and light snacks, and cleaning. Reports citing the company's announcement state that the six robots performed a total of 1,730 tasks, with four of them passing the established evaluation criteria. However, these figures are based on corporate announcements. Information regarding how many ti
Open source S4demonstration_duration
24 hours
View original evidence
uct on a hook because a shelf tag was in the way, determined that the situation was difficult to grasp with its head camera. It then switched to a procedure of checking from a closer distance using a wrist camera and successfully completed the task. ■ 24-Hour Demonstration at a Convenience Store From November 2 to 3, AGIBOT conducted a 24-hour demonstration using six "AgiBot G2" humanoid robots at a convenience store
Open source S3date
2,026
View original evidence
Can a robot apply the failures it experiences on-site to its next task after being introduced to a store? Chinese robotics company AGIBOT tested this mechanism in a retail setting. On October 27, 2026, the company announced "Genie Master," an AI foundation that allows robots to improve their movements based on their work experience. The presentation materials introduced an example where a robot, unable to hang a prod
Open source S2date
2,026
View original evidence
ed have not been confirmed. At this point, rather than being material to decide on an immediate introduction, it is a case study where we want to verify in future demonstrations "to what extent the robot can handle changes in the field on its own." ■ Sources ・AGIBOT, Genie Master Announcement (October 28, 2026) https://www.agibot.com.cn/article/315/detail/257.html ・AGIBOT, Advance Announcement of 24-Hour Convenience
Open source S8date
2,026
View original evidence
Store Demonstration (October 22, 2026) https://www.agibot.com.cn/article/315/detail/250.html ・NewAI Report on Demonstration Results (November 4, 2026) https://hub.baai.ac.cn/view/59346
Open source S9
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
uct on a hook because a shelf tag was in the way, determined that the situation was difficult to grasp with its head camera. It then switched to a procedure of checking from a closer distance using a wrist camera and successfully completed the task. ■ 24-Hour Demonstration at a Convenience Store From November 2 to 3, AGIBOT conducted a 24-hour demonstration using six "AgiBot G2" humanoid robots at a convenience store
Open source S3
in Shanghai, China. They tested multiple tasks that occur in a store, such as restocking, handing over products, serving coffee and light snacks, and cleaning. Reports citing the company's announcement state that the six robots performed a total of 1,730 tasks, with four of them passing the established evaluation criteria. However, these figures are based on corporate announcements. Information regarding how many ti
Open source S4
Can a robot apply the failures it experiences on-site to its next task after being introduced to a store? Chinese robotics company AGIBOT tested this mechanism in a retail setting. On October 27, 2026, the company announced "Genie Master," an AI foundation that allows robots to improve their movements based on their work experience. The presentation materials introduced an example where a robot, unable to hang a prod
Open source S2
ed have not been confirmed. At this point, rather than being material to decide on an immediate introduction, it is a case study where we want to verify in future demonstrations "to what extent the robot can handle changes in the field on its own." ■ Sources ・AGIBOT, Genie Master Announcement (October 28, 2026) https://www.agibot.com.cn/article/315/detail/257.html ・AGIBOT, Advance Announcement of 24-Hour Convenience
Open source S8
Store Demonstration (October 22, 2026) https://www.agibot.com.cn/article/315/detail/250.html ・NewAI Report on Demonstration Results (November 4, 2026) https://hub.baai.ac.cn/view/59346
Open source S9
Source：Robotics — Robot learning news discovery · note.com