Watch: AGIBOT puts six humanoid robots through a 24-hour convenience store shift

Chinese robotics company AGIBOT launched a 24-hour convenience-store challenge featuring six G2 humanoid robots working alongside human customers. The livestream, titled “Robots Never Clock Out,” ran from November 2 to November 3, giving viewers an extended look at the robots performing everyday retail tasks.

Their assigned duties included stocking shelves, preparing coffee and hot food, greeting customers, and clearing rubbish.

According to the firm, the challenge was designed to test the robots across different store conditions, including rush-hour crowds, overnight operations, and the morning breakfast rush.

Humanoids tackle retail

AGIBOT deployed six G2 humanoid robots in a convenience store for a 24-hour challenge designed to demonstrate their ability to perform everyday retail tasks. The event gave viewers an extended look at the robots operating in a real-world retail environment alongside human customers and store staff.

The robots are expected to handle several tasks, including stocking and organizing shelves, preparing coffee and hot food, greeting customers, and clearing rubbish. AGIBOT has divided the challenge into four operating periods: the lunch rush, evening rush, late-night duty, and breakfast preparation.

Robots Never Clock Out | AGIBOT G2’s 24-Hour Convenience Store Challenge

Six AGIBOT G2 take on a 24-hour shift as convenience store trainees. From rush-hour crowds to late-night solo duty and the morning breakfast sprint, can they keep up with nonstop customers and…

— AGIBOT (@AGIBOTofficial) November 2, 2026

The schedule was intended to expose the machines to different working conditions, from periods of heavy customer activity to overnight operations and the morning breakfast rush.

The convenience-store setting brought together several technical requirements. The G2 robots navigated spaces shared with people, interacted with merchandise, reach and manipulated objects, and responded to customer requests while other activities continued around them.

These tasks require more than basic movement. Stocking shelves involves locating merchandise, reaching for individual items, and placing them in the appropriate positions. Preparing coffee and hot food requires the robots to interact with equipment and objects in a controlled manner, while customer-facing duties involve responding to requests and operating around people.

The different operating periods also introduce variations in the environment. During rush hours, robots must work around increased customer activity and potentially obstructed routes. Overnight operations provide a different set of conditions, while breakfast preparation introduces another sequence of tasks.

AGIBOT tests autonomy

The 24-hour livestream offered a longer observation window than a short demonstration, allowing viewers to see not only successful task completion but also potential pauses, repeated attempts, and recovery from difficulties.

For example, an item may be hard to grasp, a customer may block a robot’s route, or an unexpected request may interrupt an ongoing task. How the machines respond to these situations could better indicate their practical capabilities than isolated demonstrations of individual movements.

The challenge followed AGIBOT’s AI Day announcements, which included new AI models, the G2 Max robot, and a proposed Worken measure of productive robot work. The company presents the convenience-store event as a practical demonstration of its technologies working together in an everyday environment. However, the announced 24-hour duration refers to the livestream and does not establish that each robot operates continuously without charging, maintenance, or human assistance.

The promotional material reviewed also does not specify a complete intervention policy or a scoring system for measuring task success. Experts highlighted that these details will be important when evaluating the robots’ performance. The number of tasks completed, time required for individual activities, frequency of human intervention, and ability to recover from unsuccessful attempts would provide useful measures of their effectiveness.

For now, the livestream offers an extended opportunity to observe six G2 robots performing a range of retail activities in a shared environment. It could provide a more detailed picture of how humanoid robots handle routine work beyond controlled demonstrations, although the event’s duration alone is not evidence of fully autonomous, uninterrupted operation.

Source: https://interestingengineering.com/ai-robotics/agibot-humanoid-robots-in-convenience-store