Text | Kaori

In early March 2026, Qian Xiaolei, the founder of Banana Climbing, gathered the core team dispersed across the country in Shanghai to learn about AI. He didn't want to wait any longer.

Qian Xiaolei has opened 21 climbing gyms in seven cities nationwide, with more than 200 employees. By national standards, this is a small and micro enterprise. But in the Chinese climbing industry, this is the chain brand with the most stores and the largest scale.

Over the past decade, he has transformed from a founder of an internet media and PR company into an offline industry boss who spends a lot of time each day dealing with site selection, engineering, and performance evaluations. Qian Xiaolei said these are things he didn't want to deal with but had to.

Bringing people to Shanghai was because he felt something was "off" during this year's Spring Festival. His social circle split into two groups: one consisting of peers, coaches, and members from the climbing industry; the other comprised of people he met while working in internet companies and tech media. Around the Spring Festival, the two groups exhibited extreme differentiation; peers in the climbing industry had almost no awareness of AI, while discussions among the internet group had reached a fever pitch of FOMO.

What spurred him further was the annual meeting of Tencent Research Institute. Qian Xiaolei attends every year, and the topics in recent years have been tech for good, internet ethics, environmental protection. This year, from morning until night, all topics were related to AI.

This spring, anxiety about AI technology became the main theme for this climbing gym owner.

An Urgent Offline AI Training

Qian Xiaolei stands at the intersection of two worlds. He has written about mobile phone trends on Zhongguancun Online, worked in PR at Smartisan Technology, witnessing Nokia's fall from grace and the iPhone's rise from mockery to world dominance. These experiences keep him in a state of reflexive alertness regarding technological change, but many people on his current team have never even encountered the concept of efficient work.

The meeting in Shanghai lasted two days, with Qian Xiaolei, his partners, and a soon-to-be-hired former product manager from a digital team conducting AI technology training, requiring everyone to create something with Manus based on their respective business areas.

There were no tools set up on-site, no one-on-one coding guidance, and they solved the issue of Google Mail being blocked right there. Ultimately, the conversion rate of all employees getting hands-on with AI was 50%.

What surprised Qian Xiaolei the most was a spot setter named Yang Kai.

The daily routine of a spot setter involves designing climbing routes on the rock wall, selecting locations, color coordination, and testing difficulty. It requires spatial imagination and physical experience, making it a skilled job. No one expected him to write code.

Yang Kai spent about two hours on-site creating two items.

One was a rock point inventory management system. Each store has three thousand to five thousand rock points, but once order data is lost, no one knows which points are installed in the gym. Yang Kai created a tool to label, archive, and track each rock point.

Peers opening one or two stores might stay up late to select points a couple of times and endure for two or three days, but Banana has opened twenty stores, and Yang Kai fights with Excel every month. Scale has created pain, and AI happens to emerge where the pain is most intense.

KK is the regional manager for Banana Climbing in Shanghai and Zhuhai. Before joining, he worked in a property company at a casino in Macau, managing one of Asia's most complex indoor spaces.

He used AI to create a store monitoring patrol system.

All Banana Climbing stores are equipped with comprehensive monitoring, with no blind spots except for locker rooms, as climbing involves injury risks, requiring video retrieval at any time. Since monitoring already exists, KK's idea was to let it also perform patrol functions, such as checking for trash on the floor, whether rental shoes are returned on time, if coaches are wearing uniforms, and if the front desk is using mobile phones.

The system can start inspections for all 21 stores across 7 cities with one click. Monitoring captures screenshots, AI analyzes the images, and generates reports.

This system addresses the conflict between store scale and distance. Banana Climbing's stores are dispersed in seven cities, averaging three stores per city, and traditional supervisory travel inspections incur staggering costs. This AI system's daily operational cost, including API and token fees, is only just over a hundred yuan.

However, Qian Xiaolei knows that this system can't observe everything. It can detect paper on the floor but can't smell whether the shoes are stinky. It can check if the coach is in uniform but can't sense if the venue is ventilated or if customers are feeling hot. "Sitting at the front desk, you might feel cold, but customers inside might be very hot. You must be in the customer's position, doing the same thing as them, to sense their discomfort."

To this end, Qian Xiaolei established an AI efficiency award and gave KK a bonus.

What AI can do is free people from repetitive supervision, allowing them to do things that only humans can do.

"I Am So Happy"

After the training in Shanghai, Qian Xiaolei found that the conversion rate of employees using AI was only 50%, and his initial anxiety about distance turned into technological anxiety.

Those already using AI produced results beyond expectations; however, employees who hadn't gotten hands-on, transitioning from zero foundation to proficient use of AI technology, lacked willingness, but needed someone knowledgeable in technology to help them cross the initial threshold.

Thus, Qian Xiaolei posted a recruitment ad on Xiaohongshu.

This recruitment post was not aimed at coaches but at technical individuals who understood AI and were willing to delve into specific business scenarios. People from various backgrounds responded, most of whom were climbing enthusiasts. After meeting several in Shenzhen and Shanghai, he hired two.

One of them helped his e-commerce team solve a long-standing problem.

The Tmall store of Banana Climbing, which represents outdoor equipment brands, has a vast number of SKUs, including various gear, quickdraws, mechanical belays, and harnesses, each requiring a Tmall product detail page.

The previous process involved the operations team first translating English product information into Chinese, placing it roughly in the layout file, marking the font size, and then handing it over to designers for production. The design cost for one product detail page typically ranged from 300 to 500 yuan, but the greater cost was time, as operations spent significant effort on requirements while designers spent considerable time understanding them, creating rounds of communication that could stretch for days.

The newly hired AI engineer equipped the e-commerce operations with a toolchain, connecting a few APIs, creating a custom skill that automatically generates product detail page images by inputting the product name, without any involvement from designers throughout the process.

Ultimately, the API consumed around 60 dollars, generating dozens of product detail pages. When Qian Xiaolei spoke about this, he used four words: "I am so happy."

Goats in the Valley of New York

The same thing is happening on the other side of the Pacific.

Scott is the founder of Hudson Valley Forestry, a company in the Hudson River Valley of New York, primarily engaged in land clearing, using large grinding equipment to process brush and weeds to make way for photovoltaic power stations, pipeline corridors, or private properties. They take on everything from half an acre backyards to 50-acre industrial projects.

Scott was previously a film photographer for 15 years, with no software engineering background.

In areas where machines cannot enter, or where underground high-pressure pipelines exist, he deploys a herd of goats for targeted grazing, using animals instead of machinery for clearing. He needed to track the locations of these goats; while market smart collars cost $300 each and required monthly fees, Scott created his own tracking system that cost only $30 per device.

This is not the only system he has created. He has also built the company website, CRM, and ERP, all self-hosted on his own servers. Another project he is working on is building a mesh communication network at job sites in remote valleys without mobile signal, allowing workers to send messages, share images, and exchange GPS positioning and project data.

Investor Todd Saunders used a term when sharing Scott’s case—blue-collar developer.

Qian Xiaolei and Scott are separated by 12,000 kilometers. One manages twenty-one climbing gyms and two hundred employees, while the other raises goats and operates grinders in the New York valley. Yet, what they are doing points in the same direction: in their respective industries, using AI to solve issues that previously either had to be endured, outsourced at a high cost, or were simply not seen as pressing problems at all.

But AI has arrived. Qian Xiaolei's CRM system was developed by an entrepreneur who is both a climbing enthusiast and a member of Banana Climbing.

This individual established a small company to assist coding with AI, developing and launching a member management system specifically for climbing gyms in four to five months. It includes facial recognition entry, gate linkage, and insurance integration, with all features originating from unique needs in the climbing industry. Banana is his seed customer, and he is now selling this system to more climbing gyms.

Qian Xiaolei believes that while this system may occasionally have minor bugs, its cost and the efficiency gains it offers are unimaginable compared to previously. "When the threshold and cost of writing code are low enough, I can customize; why buy standardized products?"

This statement reflects a reality currently happening in the SaaS industry of 2026. In 2024, the global SaaS industry experienced the most severe valuation contraction in a decade.

Salesforce's stock price fell over 30% from its peak in 2021, with its price-to-earnings ratio compressed from over 60 times to less than 30 times. Twilio, once viewed as a SaaS benchmark, laid off over 1500 employees in 2023. The number of fundraising rounds for vertical SaaS companies dropped by more than 40% year-on-year in 2024. Almost every star company from the last cycle has been experiencing the same story: slowing growth, increasing customer churn rates, and eroding pricing power.

The reasons are not just due to the economic cycle; in the tide of AI, those companies that once relied on generic solutions to collect subscription fees have seen their moat shift from product to inertia.

Inertia will eventually be broken.

Qian Xiaolei himself is also a practitioner of this migration; he even attempted to abolish the internal work order system. The operations team had previously implemented a work order system to manage design requests, "I was particularly opposed at the time, feeling it was a flaw only big companies have."

Later, he discovered that if someone could clearly describe a design request, it could be directly presented to AI for work without the need for an order queue. "Of course, the poster produced by AI might be a 70, but what designers create is at least an 80 or 90." The gap still exists, but for many simple requests, the 30-point difference is no longer worth the three days of waiting.

Using AI to Control the Size of the Organization

In Qian Xiaolei's climbing gym, many of the members are programmers. Some have already been laid off.

"When someone just gets laid off, they start coming to the climbing gym more frequently. They show up during the day on weekdays, and when asked, they say they've received big packages and seem very happy. After a few months, they might come asking us if we are hiring coaches."

There is a climbing friend of Qian Xiaolei's age who used to work in the internet industry, and after being laid off, went to work as a coach at a climbing gym. Qian Xiaolei thought from a different angle: he climbs well and can be a good coach, which is better than being unemployed at home.

The personal computer revolution, the internet revolution, and the mobile internet revolution in the past few decades have not managed to digitize half of the world's jobs. What AI impacts is precisely those who "won" in the earlier waves of change.

Financial analysts analyzing data in office cubicles are more at risk than construction workers moving bricks on job sites. Designers correcting images at computers are more at risk than helpers chopping vegetables in kitchens. Programmers typing on keyboards are more at risk than cleaning ladies tidying up office spaces. This is because the cost of AI replacing digital jobs is rapidly declining, while the cost of replacing physical jobs remains high.

Qian Xiaolei's climbing gym happens to be at the intersection of these two groups; his members include both white-collar and blue-collar workers. In his observation, climbing has never been an exclusive sport for white-collar workers.

Qian Xiaolei is a very optimistic person; he says, "All sad things won’t last past 2 AM; a good sleep will fix it." However, his anxiety has intensified over these two months, "As I delve deeper, I find this thing quite scary."

The fear does not lie in AI posing a fatal threat to the climbing industry; the nature of the sport is far removed from AI. The fear is rooted in his inability to imagine what large models and agents will look like in three years.