Is Bianlifeng, Which Turns Humans into Machines, Really Intelligent?
I. “You Have Been Penalised and Demoted to Shop Assistant”
A few months ago, Ms L, who served as shop manager at a Bianlifeng store in Beijing, suddenly received a message like this on her handheld mobile terminal. In her daily work, she had to keep an eye on the device—which looked much like a smartphone—every moment of the day, completing tasks that appeared on the screen every few minutes.
But this time, the notification on the screen read “Replace Shop Manager,” as mundane as being asked to restock shelves, mop the floor, or clean the display cabinets.
Ms L had been shop manager at the store since the first half of 2023. As for the reason behind her demotion, she still does not know to this day. “From start to finish, no one spoke to me about this. No one told me why. All I received was that one message from the system, and that was it—the decision was already made.”
She speculates that the reason for her demotion may have been that, a few days earlier, she had not completed the required “shop manager self-inspection” on time—checking hygiene, product placement, and other in-store conditions each day, taking photographs, and uploading them to the system.
Ms L recalls that the coffee machines and beverage dispensers, which had sat unused for a long time, were being removed from the store. She saw an opportunity to reorganise the freed-up space, displaying room-temperature mineral water and beverages to boost sales. For this project, she spent every day of that week working two extra hours of unpaid overtime on top of her 12-hour shifts, from 7 a.m. until 9 p.m.
Under such an intense workload, she “unapologetically” put the tedious self-inspection on the back burner. After all, priorities matter, and she believed everything she was doing was to increase the store’s sales.
She could not understand: she was clearly trying every way she could to boost Bianlifeng’s revenue—how could she possibly be demoted for that?
“Bringing this up truly breaks my heart,” Ms L said, describing Bianlifeng’s management as “far too dehumanising.”
Regrettably, Ms L’s grievances will never be heard by a Bianlifeng that self-styles itself as “intelligent.”
II. “The Robot Is an Idiot”
Experienced shop managers are a scarce resource in the convenience store industry. Bianlifeng claims that while traditional convenience stores need two years to train a shop manager, it takes them only six months. In founder Zhuang Chenchao’s view, it is precisely the shortage of shop managers that makes predecessors such as FamilyMart and 7-Eleven “expand too slowly”.
Bianlifeng’s proposed solution: replace shop managers’ decision-making with an algorithmic system. Zhuang Chenchao seems to sincerely believe that algorithms designed by a team of engineers in an office building can remotely understand and control every variable in the operation of all its convenience stores. In a speech, he confidently declared: “Every node involving a human being leads to a decline in overall efficiency.” The sole tasks of frontline employees are limited to “obeying system instructions and providing customers with a good service experience.”
Is the algorithm really smarter than a shop manager?
For a convenience store with limited space, achieving a precise match between supply and consumer demand is the core factor that determines a store’s revenue and profit. To demonstrate the algorithm’s superiority, Bianlifeng claims they once conducted an experiment: they selected ten experienced 7-Eleven shop managers and asked them to reduce a store’s SKUs (stock-keeping units, essentially product categories) by 10%. The result: sales dropped by 5% the next day. Using Bianlifeng’s algorithmic solution instead, sales fell by just 0.7%.
“The system even tells you in detail how many meat steamed buns and how many vegetable steamed buns to make each time, and how many boxed meals to prepare,” said Ms L. But based on her own observations, the supply arrangements dictated by the algorithm have been far from satisfactory.
“Sometimes there are far too many items and sometimes far too few. Some crisp brands clearly don’t sell, yet they keep coming in loads; some drinks sell really well, yet the next day they stop delivering them. Everyone near our store is office workers—you’d think we’d need a wider variety of bread, but we always get just the same few types, and we’re constantly complaining about it.”
Last winter, customers preferred room-temperature mineral water because of the cold, so Ms L placed some on the shelves. But the in-store patrol robot took a photograph and flagged it as products not placed in their prescribed positions, and she was ultimately forced to make corrections.
“Customers who come into the store all say: ‘Your store’s robot is so intelligent!’ But my colleagues and I sometimes genuinely can’t help calling that robot an idiot!”
Even when they consider the supply arrangements unreasonable, Ms L and her colleagues have no way of raising the issue with Bianlifeng. The system has already planned what stock to bring in and specified the exact shelf position for every product. If they have concerns about incoming stock, the only thing Ms L can do is “assign a task” to the middle platform staff within the system, for them to handle. “The process is complicated. We can barely find anyone to talk to, and this sort of small matter is no one’s responsibility.”
Despite Liu Lu repeatedly submitting requests to reduce hot meal supply, the response she received was: “Algorithmic calculations indicate that the store still has hot meal sales opportunities… The algorithm system does not yet have a relevant operational SOP (standard operating procedure), and human intervention is not possible.” As a result, as a franchisee, Liu Lu had no authority to adjust the ordering volume for any particular product in a timely manner. As the frontline operator who knows customers best, she could only watch customers drift away while bearing the operational losses caused by massive product waste.
III. “The Little Bee Must Always Be Here”
Bianlifeng’s solution: use algorithms to break down a shop assistant’s daily labour into 70–80 simple tasks, then issue instructions to staff through the handheld terminal.
Ms L completed just one week of induction training before starting work. “The main content of the training was how to operate the handheld system. As long as you learn how to use the handheld device and follow its instructions to complete tasks, you can basically handle the job. Before long, I was made shop manager.”
Ms L always tells new shop assistants: “Just get on with it!” This is the lesson she summed up from her time at Bianlifeng—don’t question, don’t overthink, because the algorithm system has already mapped out every detail of your day’s work.
“After working there for a long time, you start to feel like a walking machine. From the moment you enter the store at 7 a.m., you start preparing the hot breakfast, and once you’ve been busy until 9:30, you have to start preparing lunch.”
“We are not allowed to communicate with customers, because Bianlifeng considers talking to customers a waste of time. But customers ordering on their own are indeed more efficient—even on my own, I can serve over thirty hot meals during lunch.”
After the peak dining hours pass, shop assistants do not get to relax. Every few minutes, the handheld terminal assigns them new tasks—wiping shelves, sweeping the floor, restocking goods, conducting stocktakes, and other tedious chores. Staff are constantly prodded by tasks, kept in a state of perpetual busyness until their 7 p.m. shift ends.
“Sometimes I don’t understand what the point of wiping the shelves over and over again is. Later, I learned that their motto is simply: ‘No one is allowed to stand idle.'”
Dull and exhausting work inevitably leads to complacency among shop assistants, and that is when the “electronic overseers” come into play. In Ms L’s store, more than 30 ceiling-mounted cameras capture the staff’s every move. “When I first started, an older shop assistant told me that you just need to let the camera see your hand on the shelf—you don’t actually have to be that diligent. I didn’t dare at first, but eventually I learned to fake it.”
“Little bee, you must always be here! Little bee, you must always be here!”
IV. “Filling Labour Turnover with Labour Turnover”
Explaining her dilemma, Ms L said: “Bianlifeng actually does have its advantages. For people like us who are under a lot of financial pressure, the work tasks there are fairly simple and easy to pick up. As long as you’re willing to work hard, you can earn a decent salary each month.”
The convenience store operates 24 hours a day, with staff divided into two shifts that seamlessly overlap—12 hours each. If you’re willing, you can work 12 hours every day, 365 days a year.
In one particular month, she worked a total of 360 hours, busy from morning till night, spending each day in a daze. When she received her pay at the end of the month—a little over ten thousand yuan—she said that was what made her happiest.
Being demoted from shop manager to shop assistant was a devastating blow to Ms L. Losing the revenue-based bonus and being reduced to a shop assistant wage of 22 yuan per hour meant a sharp drop in her monthly income. As an ordinary shop assistant, she could no longer be assigned to the same store as she had been when she was manager; she could be sent by the system to any store within a 10-kilometre radius to fill gaps left by temporary leave, resignations, or dismissals.
“Most shop assistants earn 22 yuan per hour and face this extremely oppressive work environment every day. Bianlifeng’s employee turnover rate and labour mobility are extremely high.”
Ms L believes that Bianlifeng is trying to “fill labour turnover with labour turnover” to address this problem. But the instability of work locations was something Ms L could not accept, and in the end she chose not to continue working at Bianlifeng as a shop assistant.
V. When Humans Are Reduced to Machines of Flesh and Bone
Algorithms can indeed deliver some degree of “cost reduction and efficiency gains” in areas such as product selection, pricing, footfall, shrinkage rate records, and matching optimal delivery routes. Bianlifeng also relied on the concept of intelligent operations to attract investor favour, achieving breakneck expansion in its early growth phase.
However, from the end of 2021, news of layoffs and store closures began to trickle out, and in 2022 the company internally launched its “Hibernation Plan.” According to a report by Xuebao Finance, by February 2024 Bianlifeng’s store count had fallen from a peak of nearly 3,000 to just over 1,000, in stark contrast to the convenience store industry’s overall steady growth.
The abrupt pivot from “breakneck speed” to “hibernation” tells its own story: the algorithm-driven operating model has not sparked a “revolution” in the convenience store industry. At the very least, Bianlifeng’s algorithm is not as miraculous as its founder advertised or investors imagined.
As algorithms increasingly permeate every aspect of commerce and daily life, Bianlifeng’s miscalculating algorithm seems to be reminding us that arrogantly relying on algorithms while ignoring the insights and experience of frontline workers can, in fact, harm a business. More importantly, blind worship of high technology demeans human creativity itself, alienating shop assistants into numb machines of flesh and bone that simply execute algorithmic commands. In a sense, shop assistants and shop managers like Ms L are not treated by Bianlifeng as living, thinking human beings, but rather become extensions of machines and algorithms, completing tasks that machines have not yet been capable of performing.
Ms L will find her next job after leaving Bianlifeng. But beyond Bianlifeng, in manufacturing, finance, services, and even agriculture, countless algorithm-related technological innovations are continually emerging.
Amid the roar of these algorithmic machines, how can technology better serve creation rather than becoming a new mountain weighing down and alienating workers? This concerns not only the efficiency of economic operations but also the dignity of workers. Faced with such questions, we must, I fear, not only continue to pay close attention to the technology itself but also seek answers beyond it.
Unless otherwise noted, all photographs were taken by the author
Editor: Wang Hao Layout: Shi Wu