Qwen Office e-commerce edition launches as merchants exceed 20% of paying users

The e-commerce edition is positioned as an independent third-party e-commerce agent open to merchants on all e-commerce platforms. Merchant tools and industry software from different platforms can connect to it, and the product uses MCP to link directly to merchant analytics tools. According to Leiphone, Qwen Office is also rolling out proactive operations capabilities that allow the agent to detect anomalies in inventory, advertising, conversion and after-sales.

Liu Xianglong, a merchant preparing to open a 3D-printed shoe store, did not start with sampling, photography or product detail pages. He first used AI to research the market, create digital samples and prepare product images and detail pages, so that consumers could click and add items to cart before deciding whether to actually produce the shoes. Opening a new store used to take about a month from initial research to product preparation and more than 10,000 yuan in upfront investment. With tools such as Qwen Office, Liu compressed the preparation cycle to seven days and cut costs to a few hundred yuan, mainly token costs, according to the report.

Liu also runs a women's shoe store with about 430 products. Its merchant analytics tool contains 43 pages of operating data. Before using AI, operations staff had to export reports manually and search through visitor, conversion and add-to-cart metrics to find products worth investing in. He has gradually handed this work to Qwen Office and automated some repetitive operations. Through MCP, the e-commerce edition connects directly to the analytics tool. Liu converts his personal experience into prompts, generates data dashboards and sets scheduled tasks to send reports to DingTalk group chats. He now spends two to three hours a day on operations and uses the saved time to build a three-level dashboard system for platform, store and product. The platform dashboard shows daily GMV, order volume, average order value and visitor numbers for the owner.

Kan Ouli, CEO of the coffee e-commerce company Babite Coffee, faces pressure from running more than 10 e-commerce platforms with a small team. Operating data is scattered across systems, and generating a business report requires logging into back ends, exporting sales and traffic data, cleaning it, summarizing it and analyzing it. By the time the team finishes, the data is often no longer current. After connecting to Qwen Office, Kan's team can use the AI workstation to quickly aggregate store operations. During periods such as Double 11, when demand for timely data is higher, it can generate a cross-platform operating dashboard in 30 minutes.

Xiao Yu, an operations employee at the appliance company Little Bear Electric, used to check product prices on Tmall and JD every day. A single check took 15 to 30 minutes because he had to calculate coupons and platform subsidies to determine the actual price, and he could still miss short-term price cuts or delistings. During major promotions, he had to check more frequently. With Qwen Office, the system automatically collects the actual selling prices of products on the two platforms at 8:30 a.m. and 6 p.m., generates a price report and sends it to his email. The report includes changes in his own and competitors' prices, inventory status, seven-day trends and price gaps, allowing him to see anomalies in time.

Sandy, who has 15 years of e-commerce operations experience and runs two Tmall underwear flagship stores, has struggled with product creative assets. Underwear relies heavily on visual expression. In the autumn and winter new-product season, she estimated that the two stores needed about 10,000 seeding images across search, recommendation and planting scenarios, in addition to main images and detail-page assets. In the traditional manual model, a designer can make about 100 images a day, close to the production limit. Producing 10,000 images would take about 100 working days, or four to five months. Sandy said that being slow meant missing the entire peak season.

She began using Qwen Office not only as an image-generation tool but as a task scheduler. Considering the cost of large-scale image generation, she uses Qwen Office connectors to call third-party image-generation software, which organizes tasks, passes instructions and drives batch execution. The autumn and winter asset preparation that once took four months was compressed to about 15 days. Sandy said that previously the team had to wait until Christmas to get all the assets, while now it can basically finish around National Day. After automatic screening, 376 of the images generated from Sandy's instructions were usable and 72 were rejected, a usable rate of 84.0%.

AI has also changed how Sandy's team analyzes data. Its Monday data analysis once took three hours; after connecting to Qwen Office, it takes less than one hour. When facing problems such as a decline in click-through rate, the AI does not only give a conclusion but breaks the issue down from external competition, product visuals, copy and advertising structure to help find clues. Well-performing assets are placed in a premium asset pool, effective judgment standards are written into a rules library, and advertising results determine which directions to increase and which to adjust or stop. Data analysis, effect verification and experience accumulation form a loop.

As AI digital employees take over repetitive and tedious work, merchants are shifting their energy in two directions, according to Leiphone's observations. One is returning to strategic decisions. Liu has compressed his daily operations to two or three hours and uses the remaining time to judge business direction, manage AI and work on products and supply chains. Kan has begun using AI to analyze market changes and decide the direction of new coffee products. He said that no matter how deep a person's experience is, it is only the experience of his own circle, and that he now uses AI to see where the market is going before deciding whether to follow.

The other direction is managing AI and turning experience into organizational capability. Sandy said, 'AI is the hands, I am the brain.' In her view, AI can draw more than 1,000 images a day, but what counts as a good image, which scenarios should be amplified and which should be stopped are things she teaches it. She said that as a digital employee, AI can remember brand preferences, historical data and existing rules like a veteran employee, but the team must first explain its experience clearly and write it into processes. AI handles high-frequency, repetitive and standardized execution, while people handle direction and decisions, such as adjusting rules and finding industry opportunities.

The agent wave has entered e-commerce quickly partly because many merchants have incentives to use AI. E-commerce is close to transactions, and AI results can directly show up in sales, conversion rates and costs. It also builds on years of digitalization. Platforms and third-party service providers pushed merchants to adopt digital tools, management platforms and SaaS software, and some data intelligence analysis. However, the shortcomings of past digitalization have become more apparent as agents have risen in the past two years: tools on different platforms and at different stages are fragmented, and the intelligence level of vertical platforms and SaaS is limited. Many e-commerce practitioners have hesitated to choose AI tools because many products do not understand e-commerce scenarios deeply and have not connected to mainstream merchant tools.

In recent months, this situation has begun to change, according to Leiphone. Domestic AI office products are accelerating cooperation with SaaS and other ecosystem partners, trying to combine general office capabilities with vertical business experience. The Qwen Office e-commerce edition focuses further on cross-platform e-commerce operations and seeks to connect operating tools from different platforms through one AI entry point. Once tools are connected, the next problem is whether the product understands the business. Chen Yusheng said Context is particularly important. What truly understands a merchant's business is the operating data hidden in daily business activities: platform monitoring data, strategies accumulated in business documents, internal communication records and instructions exchanged with the agent. The industry long lacked infrastructure to structure massive information for agents to call. A path is emerging: first build a governable layer of enterprise context, then organize digital employees to work together on top of it.

With capabilities such as context, agents can become more proactive and do more for merchants. Qwen Office's e-commerce edition is launching proactive operations capabilities, allowing the agent to detect anomalies in inventory, advertising, conversion and after-sales. For small and medium-sized merchants that lack programming ability and staff, this lowers the threshold for using AI. Kan describes Babite Coffee as a specialty coffee e-commerce company using AI to transform itself. Before entering e-commerce, he was a musician who had never written code. Now Qwen Office, which is out-of-the-box and can be operated directly by ordinary employees, has brought AI into daily workflows and allowed him to spend more energy on business direction and strategic decisions. Sandy's two-person team uses Qwen Office to run a complete chain covering product launches, visual production, advertising operations, data analysis and strategy review, managing two stores in an orderly way.

Leiphone observed that Qwen Office's entry into e-commerce means AI operation for merchants is moving from generating a few images and writing several pieces of copy toward fuller business processes. In the past, e-commerce competition was largely a competition of execution: whoever could launch new products faster, adjust prices more frequently, produce more assets and monitor more channels was more likely to get opportunities. As AI takes over part of the execution work, efficiency gaps between merchants may widen again. Beyond execution, understanding of the industry, judgment about the market and control over products and supply chains will become more important. Kan said, 'The times are shaking so much; how can one avoid being thrown to the ground?' In his view, with AI as a technology foundation, merchants need to re-examine their ways of working and ask which tasks no longer need to be done by people.