The pair of shoes hasn’t been produced yet, but the merchant Liu Xianglong has already started thinking about how to sell them.
When planning to open a 3D printed shoes store, he didn’t rush to make samples, take photos, or create product pages right away. Instead, he let AI help him research the market and produce digital samples, and then prepared the product images and product pages to see if consumers were willing to click and add them to their cart.
In short, it’s about letting the market “vote” first, and then deciding whether to actually produce the shoes or not.
In the past, opening a new store would take a month, from preliminary research to product preparation, with initial investments exceeding 10,000 yuan. Now, with AI tools like Qianwen Office helping out, he has reduced the preparation period to 7 days, and the related costs have dropped to just a few hundred yuan—mainly due to Token costs.
This is not the first time that Liu Xianglong has used AI in business.
He also runs a women’s shoe store with approximately 430 products, and the business data in the merchant analysis tool consists of 43 pages. Before AI, operators had to manually export reports and then look for products worth investing in based on metrics such as visitors, conversions, and add-to-cart requests.
Today, these tasks have been gradually handed over to Qianwen Office, and some repetitive operational work has also been automated.
On October 14, Chen Yusen, CEO of Qianwen Office, mentioned a statistic at the release event for Qianwen Office’s e-commerce version.
In the more than two months since Qianwen Office was launched, e-commerce is one of the most deeply utilized industries, accounting for over 20% of current paid users.
Liu Xianglong is a typical user among them. From 3D printed shoes that haven’t been produced yet to the women’s shoe store he runs daily, he is thinking about the same thing: how to assign repetitive tasks to AI, reduce the cost of trial and error, save time, and focus on decisions that truly impact business performance.
Today, with the help of Qianwen Office, Liu Xianglong only needs to spend 2 to 3 hours per day handling operational work.
Three things that are most frustrating in e-commerce, now taken over by AI
Like Liu Xianglong, many businesses have similar feelings: e-commerce operations involve a lot of work. It may not seem complicated, but it is extremely exhausting.
The first thing that causes trouble for businesses is operating multiple platforms at the same time—they wish each employee could have eight hands.
Kang Ouli, CEO of 8Bit Coffee, feels this deeply. The team is not large, yet they have to manage the operations of over ten e-commerce platforms – a significant pressure.
To generate a business report, the first challenge is that business data is scattered across different systems: one needs to log in to the backend, export sales and traffic data, then clean, aggregate, and analyze it. After going through this process, the team often ends up with not the latest business information.
After joining Qianwen Office, Kan Ou Li finally didn’t have to worry about this anymore.
Especially during events like “Double 11” when there is a rush, merchants have higher requirements for data timeliness. Such AI workstations can quickly summarize the operations of various stores, generating a cross-platform dashboard in 30 minutes, thus serving as a “backing” for multi-platform operations.
The second issue that causes trouble for businesses is the need to constantly “monitor” manually, without any chance of relaxation.
Xiaoxiong Electric’s operator Xiao Yu has to repeatedly check the prices on Tmall and JD.com every day. What frustrates him most is that it takes 15 to 30 minutes each time, and he has to include coupons and platform subsidies to get the actual price he can receive. Sometimes, changes such as short-term price reductions or product removals are even missed. If there’s a promotion, he has to check more frequently.
After working at the office with thousands of questions, Xiao Yu felt much more relaxed. The system automatically collects the retail prices of products from two platforms at 8:30 am and 6:00 pm every day, generates a daily price report, and sends it to his email.
In this daily report, the price changes of domestic products and competing products, inventory status, trends over the past 7 days, and price differences are all summarized together. If there are any abnormalities, he can also see them in time.
The third issue that causes trouble for businesses is that there are too many and diverse product materials, and sometimes they can’t keep up with the peak season.
Sandy has worked in e-commerce operations for 15 years. Currently, she manages two Tmall underwear flagship stores. However, even with her extensive experience, she has encountered many difficulties with product materials.
Underwear is a category that heavily relies on visual expression. She roughly calculated that during the fall and winter launch season, in addition to the main image and product details materials, nearly 10,000 illustration designs are needed for different scenarios such as search, recommendations, and trending content. These amounts are required for two stores combined during one launch season.
In the traditional manual drawing mode, a common problem Sandy often encounters is the “time difference”.
A designer creates 100 images in a day, which is basically close to the upper limit of production capacity. 10,000 images would take approximately 100 working days, and it would take four or five months in total.
But time waits for no one. In her words: “Slow is like missing out on the entire peak season.”
Later, she started using Qianwen Office: not just as a generative image tool, but also for task scheduling. Considering the cost of generating large amounts of images, it called third-party image generation software through Qianwen Office’s connector, which organized tasks, transmitted instructions, and facilitated batch execution.
Sandy no longer has to be stuck in the old “traditional” method of producing and submitting images item by item. The preparation of autumn and winter materials, which used to take four months, has now been reduced to about 15 days.
“Previously, I had to wait until Christmas to get all the materials, but now I can finish it around National Day.” Sandy finally manages to prepare the materials before the peak season arrives.
The plant promotion images generated by Qianwen Office based on Sandy instructions, after automatic filtering, resulted in 376 available/72 eliminated, with an availability rate of 84.0%
As AI becomes more capable, what else can businesses do?
After AI deeply integrates into merchants' daily operations, what changes is not just the resolution of certain pain points or the improvement of certain processes, but also the reorganization of how information, tools, and human resources are distributed.
For example, many e-commerce professionals are experiencing such a change: in the past, people had to find data themselves, while now AI can actively organize information and synchronize business operations.
Facing a business analysis tool with 43 pages at a time, Liu Xianglong often felt at a loss and didn’t even have the patience to go through each page one by one.
This trouble was alleviated as Qianwen Office E-commerce version connected directly with merchants analysis tools via MCP.
Previously, it was a purely manual task that required manually downloading Excel files for analysis and relied heavily on human experience. Now, it can be completed semi-automatically: he transforms his personal experience into prompts, generates data dashboards, and can also set scheduled tasks to send reports to a DingTalk group chat.
He also stepped away from the cumbersome basic work, and used the saved energy to create a three-level dashboard system at the platform, store, and product levels by hand, so as to do more creative things.
In the three-level dashboard system, the “Platform” dashboard displays key metrics such as daily GMV, order volume, average customer value, and number of visitors for the entire store, for the owner to review.
In the perspective of e-commerce professionals, The value brought by AI is not just about speeding up processes and reducing burden, but also about connecting the various scattered business steps together.
Previously, Sandy’s team spent three hours a week on data analysis every Monday morning. After integrating Qianwen Office, this time was reduced to less than an hour.
More importantly, when facing issues such as a decline in click-through rates, AI doesn’t just come up with a conclusion. Instead, it analyzes the situation from several perspectives including external competition, product visuals, copywriting information, and placement structure, helping her find clues.
After AI analysis, high-performing content will be placed in the high-quality content pool; effective judgment criteria will be written into the rule library; then, combined with the placement results, it will be determined which directions should be intensified, which should be adjusted or stopped.
Data analysis, effect validation, experience accumulation—these are connected in a chain, turning around.
When AI digital employees begin taking over repetitive and tedious tasks, where should businesses focus their efforts?
According to Leifeng.com's observations, there are mainly two aspects:
First, return to strategic decision-making and focus on more important matters.
Liu Xianglong has now compressed the entire day’s operations into two or three hours. The time saved is used to determine business directions, manage AI, and oversee products and supply chains.
Kan Oli also starts using AI to analyze market changes, and then decides on the development direction for new coffee products.
“No matter how deep a person’s experience is, it is still within their own circle. In the past, when creating new products, I decided what to sell. Now, using AI, I first see where the market is heading, and then decide whether to follow that trend or not.”
Second is managing AI, turning experience into organizational capabilities.
"AI is the hand, I am the brain." Sandy believes that AI can draw more than a thousand images a day, "but what constitutes a good image, which scenes should be enlarged, and which should be stopped, is something I teach it."
In her view, as digital employees, AI can remember brand preferences, historical data, and existing rules just like traditional employees. The key is for the team to clearly explain their experience and write it into the processes.
These businesses that deeply use AI have clearly divided their tasks.
As Sandy said: “AI is responsible for high-frequency, repetitive, and standardized tasks, while humans are in charge of direction and decision-making—such as adjusting rules and identifying industry opportunities. This is the foundation on which the human-machine collaboration model can continue to function.”
Who will benefit first as the Agent trend enters e-commerce?
“AI will not replace humans, but it will replace those who cannot use AI.” In Liu Xianglong’s view, Qianwen Office is not just an AI tool; it is more like a “reliable e-commerce operation partner”.
In the e-commerce industry, there are many pioneers like Liu Xianglong. Their daily operations are being reshaped by agents.
The Agent Wave can quickly enter the e-commerce field on one hand because most merchants have a driving force to use AI.
The e-commerce industry is close to transactions, and the effects of AI can be directly reflected in sales volume, conversion rate, and costs. Naturally, businesses are willing to try more.
On the other hand, it cannot be separated from the industry’s years of digitalization accumulation.
From the perspective of digital infrastructure, the e-commerce industry was already digitized a long time ago by the tools provided by major platforms and third-party service providers—such as merchant management platforms, SaaS software, and certain data intelligent analysis capabilities. The foundation is not weak.
But the shortcomings left by digitalization in the past have gradually become more pronounced since the rise of agents in the past two years: the tools on different platforms and at various stages are isolated from each other, and the level of intelligence for vertical platforms and SaaS is also limited.
That’s why in the past, many e-commerce professionals hesitated when choosing AI tools, and sometimes couldn’t even start. Their concerns were mainly due to the fact that most products didn’t have a deep understanding of e-commerce industry scenarios, and many failed to integrate with mainstream business tools.
Leifeng Net (Public Account: Leifeng Net) has observed that in recent months, this situation has begun to change:
Domestic AI office products are accelerating cooperation with ecosystem partners such as SaaS, trying to combine general office capabilities with business experience in specific vertical fields.
The Qianwen Office E-commerce Version further focuses on cross-platform e-commerce operations, attempting to connect operating tools across different platforms through an AI portal.
The Qianwen Office E-commerce Version is positioned as an independent third-party e-commerce Agent. It is open to all e-commerce platforms and merchants in the market, and merchant tools and industry software from various platforms can also be integrated.
The tool has been connected. Next, what needs to be addressed is “understanding the business”. The level of intelligence can be improved, thanks to industry efforts in exploring “context”.
In Chen Yusen’s view, Context is particularly important. Those who truly understand a business’s operations are the operational data hidden in daily business activities—platform monitoring data, strategies stored in business documents, internal communication records, and instructions exchanged with Agents.
Previously, the industry has long lacked infrastructure that can structure massive amounts of information and be used by agents. As a result, the industry has gradually developed a approach: first, establish a manageable corporate context, and then organize digital employees to work together on top of it.
With basic capabilities such as context, the Agent will also be more proactive and able to help businesses do more.
It is understood that the Qianwen Office e-commerce version is launching proactive management capabilities. The Agent can actively detect abnormalities in aspects such as inventory, placement, conversion, and after-sales services.
For small and medium businesses with limited programming skills and a limited workforce, this undoubtedly reduces the barrier to using AI.
Kan Ouli designated 8bi as “using AI to transform his premium coffee e-commerce company”. Before entering the e-commerce industry, he was a musician who never wrote code. Now, Qianwen Office, which is ready to use right out of the box and can be operated by ordinary employees directly, has brought AI into his daily work routine, allowing him to spare more energy for strategic decision-making and business direction.
Sandy led a small team of two, using Qianwen Office to complete the entire process of product launch, visual production, placement operations, data analysis, and strategy review, keeping both stores well-organized.
Conclusion
Qianwen Office has entered the e-commerce market. At least for businesses, this means the AI integration in e-commerce operations, moving from just generating some images and writing some copy to more comprehensive business processes.
In the past, e-commerce competition was largely a competition of execution capabilities: whoever could launch new products faster, adjust prices more frequently, produce more content, and target more channels, would be more likely to gain opportunities.
When AI begins to take over some of the execution work, the efficiency gap between businesses may also widen again. Beyond execution, understanding the industry, making judgments about the market, and controlling products and supply chains will become even more important.
"With the times being so unstable, how can one not be thrown onto the ground?"
In Kan Ou Li's view, with the technical foundation of AI, businesses need to re-examine their working methods: Some tasks can actually no longer be done by “people.”
Leifeng Net original article, unauthorized reproduction is prohibited. For details, see Reproduction Guidelines.