Listening to the rain, from Aofeisi Temple
QuantumBit | Official Account QbitAI
Opening a 3D printing shoe store used to take 30 days, now it takes only 7 days.
Liu Xianglong, who has been in e-commerce for nearly ten years, just opened a new store like this.
He is one of the first users of Qianwen Office, and entrusted the market research and product detail page creation before opening the store to AI for assistance.
According to him, in the past, to conduct research, we had to purchase services from consulting agencies. For new product launches, we also needed to make samples and take photos, which would cost tens of thousands of yuan per round.
This time, he used Qianwen Office to complete the preliminary research and material preparation, and the Token fee cost only a few hundred yuan.
And indeed, those who have opened stores know that the greater risk is when the goods are ready but cannot be sold.
Once new products go out of stock, the money invested remains trapped in inventory, and funds don’t come back for a long time. The subsequent turnover becomes also difficult.
Now, Liu Xianglong first uses Qianwen Office to conduct research and generate product detail pages, allowing for sample testing on the same day. Once the data is available, he will decide what to produce and in what quantities.
(Test: A small-scale test of a product’s market response)
Liu Xianglong said that if not for Qianwen Office, he wouldn’t have started operating so quickly.
When we consider opening a store again, maybe others have already started doing it.
What? You say Qianwen Office can also help people open a store at any time?
Just now, Qianwen Office launched the e-commerce version, turning the product selection research, material creation, and business analysis that merchants did on their own into professional features.
The biggest feature is designed specifically for the actual business needs of merchants.
Chen Yusen, CEO of Qianwen Office, said: “The e-commerce version of Qianwen Office is positioned as an independent third-party e-commerce Agent, open to all e-commerce platforms and merchants in the market.”
Moreover, Qianwen Office also stated that it will integrate merchant tools and industry software from more platforms.
Thus, Qianwen Office also answered a larger question:
What should the Agent look like when it moves from a personal assistant to enterprise productivity?
Four core functions, integrating industry needs into products
Liu Xianglong uses Qianwen Office for market research, with the core goal of determining whether a new business can be launched.
He needs to understand what consumers want to buy, and also consider the cost of testing designs, so as not to keep money in inventory from the beginning.
Such tasks obviously go beyond the regular work of a personal assistant.
Popular personal agents today, such as Muse, provide assistance based on individual work habits and tasks.
Organizing emails, arranging trips, handling daily affairs... Its product form is more like “a personal assistant”.
But when entering an enterprise, the Agent must also become familiar with how the business operates.
For example, when operating an business wants to decide whether to add more marketing efforts for a product, it is necessary to refer to the financial costs and profits. The Agent needs to be involved in this process, which requires the context of the enterprise.
And Qianwen Office has previously integrated with DingTalk, allowing it to access group chat and document information within the enterprise’s authorization scope. The Agent can clearly know what the team has been doing recently and any related discussions.
Entering the e-commerce industry, agents also need to further understand business information such as products, inventory, and placement.
For example, which selling points should the new products highlight, what brand style should the materials match, and how to investigate abnormal issues after deployment – all of these need to be specifically designed according to the merchant’s working methods.
Today, the new Qianwen Office E-commerce Version is available. Its four professional functions are centered around this business process.
Trend Insights and Selling Point Analysis
Before releasing a new product, merchants often need to spend time evaluating the market.
For example, what outfit trends are popular on TikTok and REDnote recently? What pain points and needs do people have? All these may affect the selling points and content direction of products.
“Trend Insights and Feature Analysis” can help find clues from popular keywords and content on these platforms, providing merchants with directions for promotion.
Batch generation of popular content materials
With a direction, the next step is to prepare the materials.
The same selling point needs to be presented in different image or video scenarios; when changing channels, the way the content is presented also has to be adjusted.
And the batch generation also has to meet brand requirements; each image cannot have its own style.
“Batch generation of popular content” allows brand color schemes, visual styles, and product features to be included in the generation requirements, enabling the production of content suitable for various categories and channels in bulk.
Cross-platform combat interface
After the product starts being sold, the merchant will face a bunch of scattered business data.
The sales of a store have increased. Is the inventory able to keep up? The promotion has brought in traffic. Has the refund amount also increased?
When managing multiple platforms and stores, operators often have to open each backend separately, and then compare the reports together.
“Cross-platform operation platform” can aggregate these information, analyze sales and traffic in relation to each other, and incorporate inventory, refunds, and other aspects into a comprehensive business overview.
Who could understand? There’s no need to open each backend and compare report data anymore!
The merchant can fully focus on checking for abnormalities, and then decide which task to handle first～
Cross-platform ROI intelligent optimization
The deployment also requires continuous follow-up, as the budget is consumed daily, and it is difficult for a single analysis report to cover subsequent changes.
“Cross-platform ROI intelligent optimization” retrieves reporting data through scheduled tasks, alerts when ROI drops or budget consumption is abnormal, and provides adjustment suggestions.
Merchants can use this information to decide whether to continue observing or adjust the materials and placement arrangements.
From product selection to launch, the four major functions of Qianwen Office e-commerce version are designed around the business needs of merchants, applying general analysis and generation capabilities to specific operational processes.
For the Agent to play a role in these processes, it is necessary to deeply understand the business logic of the merchant. This also indicates that industryization is the inevitable path for enterprise-level Agent value realization.
Only by having a deep understanding of how a business operates can an Agent truly participate and enhance industry productivity.
Add the existing software, prepare 15 days of material for 4 months
Seeing this, you might wonder:
My store already has an ERP and uses many business software. How should these be used together when integrating Qianwen Office?
Normally, merchants check business data in tools such as Business参谋, but inventory and financial information may be stored in other systems. To determine exactly how much money a product generates, it is often necessary to download multiple reports for organization.
This issue has already been resolved in the Qianwen Office E-commerce version.
It can call these existing tools through connectors, organizing data extraction, processing, and analysis into a workflow.
Eightbit Coffee is a small company with only 17 employees. Its business covers more than ten e-commerce platforms, and it used DingTalk for daily approvals and financial collaboration.
After accessing Qianwen Office, they continued to use these tools and fed authorized data into the workflow via connectors.
In the past, the team’s development of new products mainly relied on the founder’s personal preferences and interactions with peers; now, with Qianwen Office, the team has established a Product Selection Research Skill.
Qianwen Office first reviews its own products, identifies those with search traffic but low conversions, then creates a competitor pool to supplement the product detail page information.
When analyzing a cold-drunk coffee, the report will compare the processing methods and flavors of different products, along with price performance, to provide formula references for the team.
Eight Bit used sun-dried beans less frequently in cold extraction recipes in the past, but with this research Skill, the team found that among similar products analyzed, sun-dried beans ranked first in processing statistics, while the washing method preferred by the team only came in third place.
The team adjusted the recipe accordingly, and combined it with the supply chain and actual taste to launch a new cold brew version of瑰夏.
The effect is immediate. According to the merchant, the sales of this new product are about 10 times that of similar products in the past.
The same approach is also used in the weekly reports and profit calculations:
The data that was originally required to be logged in manually to the backend and downloaded and organized can now be obtained and analyzed by Qianwen Office according to set cycles, and the results can also be pushed to the DingTalk group. After receiving the weekly report, the team can directly discuss business fluctuations, reducing the previous preparation work.
In addition, 8bit also connects to Juhuitan ERP and Shangshi Shoupan through the MCP connector. Data is exported and cleaned up by Qianwen Office, and then combined with the Dingding intelligent finance application to build an analysis suite.
Fees from different systems are combined for calculation, allowing the team to see the net gross margin for each channel.
Sandy's underwear team is also a typical example. The team has only two people, and they operate two brands. During a new product launch season, they need to prepare nearly 10,000 marketing images.
The team integrated third-party image editing software into Qianwen Office, and the Agent is responsible for batch generating images and organizing materials. According to Sandy, the preparation period for autumn and winter materials has been reduced from 4 months to about 15 days.
Business analysis is also integrated into the workflow. Qianwen Office reads data from the business system, organizes it, and generates analysis reports. The time required for individual data analysis is compressed from about 3 hours to under 1 hour, enabling the team to conduct subsequent deployment reviews.
In these tasks, ERP and business analysis tools continue to provide their respective professional capabilities, while the Agent calls up tools based on the merchant’s goals, turning steps that would otherwise require manual coordination into a cross-system task.
Dispersed data can be used together for product selection or profit calculation, and merchants also reduce the work of repetitive calculations and organization.
The Agent becomes a new intelligent layer on top of industry software, enabling the capabilities of existing tools to be applied to more complete business processes and creating incremental value for existing software.
This connection and collaboration also provide a new way of cooperation between Agents and SaaS.
The next competition for enterprise Agents in the industry
Why does Qianwen Office start with e-commerce?
Because e-commerce is a starting point where effects are easily seen.
The process of a product from listing to sale is mostly completed online. Business data is continuously generated, and the corresponding software is also quite mature. The Agent has the conditions to directly participate in actual work.
How long did it take to prepare the materials? How much labor was saved for a report? Businesses can directly compare the results before and after.
Accordingly, small and medium-sized businesses also need Qianwen Office.
Teams like Eightbit have few people, but operate many platforms. It’s difficult to assign a dedicated person for every task. For them, daily work already takes up most of the time, and new product development still needs continuous follow-up with the market.
Even if there are various software programs available, there is a lack of specialized technical personnel to connect these scattered tools and data together.
The significance of Qianwen Office lies in enabling small teams that lack technical support and have limited personnel to effectively use existing software and accomplish more business tasks.
In terms of the technical foundation, Qianwen Office connects with DingTalk to obtain enterprise context, which helps the Agent get the information needed for tasks and reduces the work that employees have to redo each time they interpret the business.
The self-developed Agent framework it uses breaks down complex requirements into multiple steps, and then selects the appropriate tools for execution.
For example, when creating a business report, after retrieving the data, it is still necessary to clean and analyze it. The agent needs to remember the statistical period and analysis goals, as well as know which steps have been completed, so that it can continue to use tools to handle the remaining work.
The context compression and session recovery design of Qianwen Office is used to support such long tasks: retaining the information needed for subsequent execution, allowing the task to continue progress across multiple tool calls, and enabling recovery after interruption.
For businesses, these engineering capabilities are not only related to whether a report can be completed as required, but also relate to the operational tasks that are carried out repeatedly daily, and how much labor costs can be saved.
It can be said that the Qianwen Office e-commerce version has set a example for enterprise agents:
Connect existing tools in the industry, and design features around actual business operations, so that general office agents can be integrated into specific business processes.
As an independent third-party Agent, it is open to the entire industry ecosystem and will continue to access more platforms and tools.
The release of the e-commerce version also sent an industry signal: Competition in enterprise-level Agents is moving beyond model capabilities to include industry understanding and ecosystem connection.
What tasks the model can perform must ultimately be tested in an enterprise’s operations.
Whoever can truly connect industry tools and understand industry logic will be able to define the production and management methods of the next generation of enterprises.
In the future, Qianwen Office E-commerce version also plans to expand into more industries, continuing to enter fields such as finance, government affairs, and education.