7-Eleven Tests Store and Delivery Robots as Japan's Worker Shortage Deepens

Robots take on the daily work of keeping stores open

Seven-Eleven Japan is testing robots at a Tokyo convenience store for about three months, targeting chores that consume hours of employee time. A machine stocks bottled drinks and canned alcohol, cleaning robots handle floors and windows, and a screen near the register connects shoppers with a remote customer support agent. The company estimates that beverage restocking alone takes employees one to two hours a day.

The retailer wants to find out whether this combination can help keep stores operating around the clock as workers become harder to recruit. If the trial delivers strong results, it plans to extend the approach to more locations, aiming to reduce employee workload by roughly 30 percent. The time saved would go toward customer service, fresh food preparation, and developing products and services.

As of September 12, 2025, the store experiment is part of a wider push into automation. Seven-Eleven is also testing delivery robots in Hachioji, west of central Tokyo, in an area covering roughly 10,000 households. That trial is scheduled to continue into early 2026.

Japan's convenience stores, known as konbini, provide much more than snacks. They sell hot meals and offer services such as bill payment, making them important stops in everyday community life. Maintaining that service through the night has become harder as hiring dries up and labor costs rise. Robots are moving from storage areas into shopping aisles and onto neighborhood sidewalks to help fill those gaps.

Why the labor shortage is driving the trials

Around a third of Japan's population is now over 65, while its working age population has been shrinking since the late 1990s. That leaves businesses competing for a smaller pool of people to staff stores, prepare food and deliver goods.

A 2023 study by Recruit Works Institute projected a shortage of about 3.4 million workers by 2030 and roughly 11 million by 2040. Those figures describe a growing gap between the workers needed and the workers available. The projected shortage in 2040 is more than three times the figure for 2030.

Japan already uses industrial robots extensively. The International Federation of Robotics ranks it fifth worldwide for robot density in manufacturing, with 419 robots per 10,000 employees. Service businesses are the next area of attention because they account for most jobs and face the most acute shortages.

Policymakers and executives increasingly describe automation as a way to maintain output when there are too few workers. Analysts at the International Monetary Fund have examined Japan's history of robot adoption and found that, in many manufacturing settings, automation coincided with higher productivity and steady employment.

Productivity means how much work can be completed with a given amount of time and resources. Japan's service sector has lagged on that measure for years. Convenience stores offer a practical testing ground: they operate nonstop, repeat many of the same tasks every day and struggle to fill shifts. Restocking a refrigerator or cleaning a floor gives a machine a specific job whose results managers can measure.

Inside the Tokyo store trial

The Tokyo pilot divides work between machines built for different tasks. A compact stocking unit moves along chilled cases, uses a robotic arm to select bottles and cans, and places them in orderly rows. Its assignment includes bottled drinks and canned alcohol.

Beverage replenishment is a deliberate starting point. Employees must move products from storage into refrigerated displays, often in cramped, cold spaces. Busy stores need that work repeated several times a day. Taking it off the staff schedule could release one to two hours daily for serving shoppers, preparing food or tidying shelves elsewhere in the store.

Cleaning units sweep and mop floors and polish windows when aisle traffic is quieter. These chores can be postponed when a store has too few employees. Automating them is intended to keep routine maintenance going while staff deal with customers and other work.

The remote customer service station addresses another pressure point. Shoppers can use a screen near the register to contact a live agent working away from the store. The connection offers support during quiet hours, when fewer employees are scheduled, or when a cashier is on a break. The aim is consistent service without requiring a full team on site every night.

Hiroki Takei, head of operations at 7-Eleven, described the purpose of the trial in a statement shared with local media:

We aim to increase productivity and create an environment where we can challenge ourselves to create new product assortments and services.

Company managers will track labor time saved, how often drinks are out of stock and how shoppers respond. If the results support expansion, busy urban stores with tight staffing and heavy beverage restocking demands could be among the first locations considered.

How a robot knows what to put on a shelf

Shelf stocking robots are designed around a narrow, repeated task. Cameras scan the refrigerated display and compare it with a planogram, a layout showing where each product belongs. The system identifies empty spaces and ranks what needs replenishing, using factors such as past sales and the time of day.

The machine then travels along a guide rail, reaches into a storage bin and grips a bottle or can at a calculated point. It places the item in the correct shelf position. A fallen product, a blocked storage bay or another unusual situation can trigger an alert to a human overseer.

Systems of this kind typically use several computer vision models at once. One identifies products, another estimates depth so the arm can reach accurately, a third checks for gaps or misplaced items, and a fourth looks for unusual conditions. Computer vision is software that interprets images, allowing the machine to turn a camera view into instructions for movement.

Developers train these models using real store footage and synthetic images created in simulation. Compact computers inside the robot, called edge computers, allow it to make immediate decisions even when a network connection is slow. Sales history and inventory targets can arrive through cloud systems to help set restocking priorities and reduce empty shelves.

Routine steps run automatically, while people handle exceptions. That combination is intended to keep errors low without requiring someone to supervise the machine continuously from inside the store. It also means automation still depends on human support when the environment becomes difficult to interpret.

Hachioji tests delivery at walking speed

Seven-Eleven's Hachioji experiment moves the task outside the store. Customers order through the 7NOW app, employees load purchases into a delivery robot, and the machine travels along sidewalks at roughly 6 kilometers per hour. The target delivery window is about 20 minutes.

The robot follows traffic signals and avoids obstacles on its route. When it reaches the customer, a QR code scan opens the compartment so the purchase can be collected. Remote supervisors monitor the journey and can take control when needed.

The retailer chose a hilly testing area to examine how the machines cope with slopes and to gather feedback from older residents. Hills and heavy shopping bags can make even a nearby store difficult to reach. A reliable delivery service could make daily essentials easier to obtain for those residents.

A company spokesperson explained why the location was selected:

We deliberately targeted an area with slopes and an aging population to clarify the problems we would face ahead of putting the service into practical use.

The test also addresses Japan's shortage of truck drivers, which has put pressure on home delivery and last mile logistics. Last mile logistics means the final part of a delivery journey, from a nearby store or distribution point to the customer's door. In this trial, store employees still prepare and load the order, while the robot carries it through the neighborhood.

The store and sidewalk trials address different parts of the same staffing problem. One aims to reduce repetitive work inside a shop; the other tests whether a small machine can handle a short delivery trip. Together, they examine how convenience stores might continue serving customers with fewer workers available.

What keeps sidewalk robots safe

Delivery robots use lidar and cameras to identify curbs, pedestrians, pets and other objects. Lidar measures distances using light, helping the machine understand the space around it. Those readings support route planning, obstacle avoidance and movement through crossings in response to traffic signals.

Sensors also watch for sudden movement. If a child darts in front of a robot, it stops. Speed limits keep travel at walking pace, and routes are geofenced, meaning the machines operate within a defined area rather than roaming freely.

A remote monitoring hub provides human backup if a route is blocked or the robot encounters a confusing situation. Low speeds, restricted operating areas and the ability to intervene are intended to reduce risk while the technology develops.

Japan changed its rules in 2023 to allow low speed delivery robots on public roads under specific conditions. Requirements include remote monitoring, audible alerts and strict speed caps. The change opened the way for trials in real neighborhoods rather than limiting them to private campuses.

Cities and police departments coordinate with retailers on test routes, crossings and operating hours. The framework resembles rules for slow mobility devices and focuses on reducing conflicts with pedestrians. Safety will remain a practical test of how well robots share space with people on everyday journeys.

FamilyMart and Lawson pursue their own systems

FamilyMart has worked with Tokyo startup Telexistence to deploy autonomous restocking systems in hundreds of stores. Its TX SCARA model operates inside beverage cases, moving along a rail to take bottles and cans from storage and place them on shelves.

The system can replenish up to 1,000 items a day and operate around the clock. An artificial intelligence platform examines shelf images and sales data to decide which products need attention first. Depth and color sensors guide the arm as it moves each item.

When a problem occurs, a remote operator can switch to teleoperation through a virtual reality console and correct it within seconds. Teleoperation means a person controls the machine from another location. It gives the system a way to deal with situations its automatic software cannot handle on its own.

FamilyMart pays through a robot as a service subscription rather than buying the machines outright. The arrangement makes access to the equipment a recurring service expense instead of requiring the retailer to purchase each unit.

Lawson has also tested shelf stocking robots controlled remotely and explored expanding trials as systems become more autonomous. Remote operation allows developers to collect training data and handle unusual cases while stores gain staffing flexibility. Operators can log in from anywhere, creating opportunities for people who cannot commute or who prefer flexible hours.

Companies in e commerce and electronics have tested delivery robots too. The activity in 2025 follows years of gradual experimentation, during which machines have moved beyond novelty displays toward specific, repeatable jobs. FamilyMart's deployment already shows beverage replenishment being automated across hundreds of stores; Seven-Eleven is testing a broader combination of stocking, cleaning and remote customer support.

What changes for employees and shoppers

Japan's demographic pressures shape the employment debate. Many executives are trying to cover shifts they cannot fill and preserve services as the working age population declines. Stores still need stocked refrigerators, clean floors and people able to help customers.

Robots could allow employees to spend more time on tasks requiring judgment and empathy, including helping an elderly shopper, handling a return or developing a local menu. Seven-Eleven's workload reduction target links those everyday benefits to the time machines might save.

Niki Harada, an official at Japan's Restaurant Workers Union, described the uncertainty facing service employees:

It is difficult to tell now what impact robots might have in restaurants, it could mean fewer people, but it could also create new jobs.

Customers may notice the changes most during quiet hours. A remote agent on a screen might help with a complex payment or age verification while a cleaning robot works nearby. The intended benefits are fewer empty shelves, consistent service and faster access to deliveries.

Older residents in Hachioji are a particular focus because slopes and heavy bags can make shopping difficult. If delivery robots prove reliable, nearby convenience stores could become small hubs for delivering essentials on the same day. The usefulness of that service will depend on how well the machines cope with the actual streets and customers they are meant to serve.

Costs, privacy and the decisions still to come

Sidewalk robots must deal with narrow spaces, difficult weather and occasional vandalism. Their vision systems also capture images of public places, making privacy part of the deployment debate. Retailers and municipalities will need clear rules covering how long data is kept, incident reporting and insurance.

Low speeds and human backup are central to the current approach. Operators will also need to demonstrate safe performance as the number of machines grows. A small trial gives managers a setting in which to examine problems before committing to a larger fleet.

Seven-Eleven will evaluate the store experiment over about three months. Its decisions will depend on reliability, shoppers' reactions and whether the machines consistently reduce workload while maintaining service quality. No date for a wider rollout has been given.

The Hachioji delivery pilot is scheduled through early 2026. Executives will compare reliability, customer uptake and the cost of each delivery with wage and fuel prices. Hardware costs tend to fall as production increases, while software can improve as systems learn from more operating hours. Expansion will depend on whether those gains translate into dependable, economical service.

The main dates and testing periods are:

• 2023: Japan allowed low speed delivery robots on public roads under conditions covering monitoring, alerts and speed.
• September 12, 2025: Seven-Eleven's Tokyo store experiment and Hachioji delivery trial form part of its effort to address staffing shortages.
• About three months: The planned evaluation period for the Tokyo store trial before a decision on expansion.
• Early 2026: The scheduled end period for the Hachioji delivery test.

What to Know

• Seven-Eleven Japan is testing drink stocking robots, cleaning machines and remote customer support at a Tokyo store.
• Beverage restocking takes employees an estimated one to two hours daily. The company aims to reduce workload by roughly 30 percent if the approach expands.
• The Hachioji delivery trial covers about 10,000 households, uses the 7NOW app and targets delivery in about 20 minutes.
• Delivery robots travel at roughly 6 kilometers per hour, with sensors, defined routes and remote human supervision.
• Japan's projected worker shortage rises from about 3.4 million in 2030 to roughly 11 million in 2040.
• FamilyMart has deployed autonomous beverage stockers in hundreds of stores, while Lawson has tested remotely controlled shelf stocking.
• The store trial lasts about three months, and the Hachioji test is scheduled through early 2026. Expansion depends on cost, reliability, safety and customer acceptance.