Business judgment on AI products
There is public usage data. The useful view is what it does, its scale, and whether the change lasts.
01
Start inside the user's day · Public facts + observable behavior · 2026-08-28
Use 千问 in the browser, with typing and meeting-notes tools sitting beside it. Monthly visits: 44.5 million.
Public materials do not yet show how users complete this job today or what they replace.
The product targets friction in this job, but public user evidence does not yet show the cost, frequency, or consequence of leaving it unsolved.
A giant opening several doors for one model is testing which door people actually stay in. Don't clone another general chat—own the tiny high-frequency desktop jobs: look-up, meeting notes, typing. Make them always-on, not a destination.
A public record shows 44.53M monthly visits and 0.56% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
① Whether web visits leave the flatline within three months — new features like the office assistant and IME, if they create increment, will show up here first; ② Whether Alibaba keeps disclosing the 295M cross-platform MAU (March figure) next quarter and its growth — the merged scope is more meanin…
Worth dissecting. A public record shows 44.53M monthly visits and 0.56% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
copy the division of roles across multiple doors of one brand — growth door (app/new-user), retention door (web/office), high-frequency touchpoints (IME/selection), all strung on the same model base. Even a small team can build the high-frequency small action as a lightweight client first, which beats building a second chat page. none independent. The free strategy is subsidized by the parent ecosystem and not directly transferable to standalone products.
Mostly free to consumers (consistent with Alibaba's stated Qwen strategy); the revenue lands in the Alibaba Cloud ecosystem: in fiscal year 2026 (through the March 2026 quarter), Alibaba Cloud AI-related product revenue was 8.971B RMB, elev… ① Whether web visits leave the flatline within three months — new features like the office assistant and IME, if they create increment, will show up here first; ② Whether Alibaba keeps disclosing the 295M cross-platform MAU (March figure) next quarter and its growth — the merged scope is more meanin…
Use 千问 in the browser, with typing and meeting-notes tools sitting beside it. Monthly visits: 44.5 million.
A public record shows 44.53M monthly visits and 0.56% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
An English validation note will follow from the public evidence.
The product claims to help users complete: “Use 千问 in the browser, with typing and meeting-notes tools sitting beside it.”. User evidence has not yet verified pain intensity or the cost of doing without it.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
02
Market comparison
The Chinese–English market comparison is not complete yet. A conclusion follows only after its coverage and verifiable evidence are recorded.
03
The call and next move come first; the full read retains the evidence and counterevidence.
Alibaba Qwen's web/PC entry point at qianwen.com — the same model as the Qwen mobile app, but through a browser and desktop front door, plus lightweight clients (Qwen IME, meeting notes, text-selection tools) that turn Qwen into a set of efficiency tools "always within reach."
Alibaba; Qwen is the foundation-model brand of Alibaba Cloud. qianwen.com and the "千问 - 阿里最强模型官方AI助手" app are two front doors for the same product — qianwen.com handles web and desktop, the app handles mobile. traffic board counts them as two separate products; merged, they are the full picture of Qwen.
Read: a giant opening multiple doors for the same model is a test of which brand asset and which usage scenario actually carries traffic. The web door's value is serving office scenarios and PC users; the app is the new-user engine. qianwen.com's MoM is nearly flat (+0.56%), which means the web door is not a growth engine — it is the stable base.
For PC users, using AI used to mean either opening a browser to various model websites or installing a collection of single-purpose tools — an IME, a recording transcriber, a selection translator, each separate. The qianwen.com bundle replaces that scattered state of "AI chat + IME + recording transcription + selection translation" across multiple apps, making AI a persistent layer you can invoke on the desktop at any moment.
Read: this door is shaped more like an office suite than a chat assistant — IME, meeting notes, and selection tools are high-frequency, low-threshold actions that build muscle memory better than a bare dialog box. It is a stake Qwen plants in the office scenario, not the main battlefield.
Mostly free to consumers (consistent with Alibaba's stated Qwen strategy); the revenue lands in the Alibaba Cloud ecosystem: in fiscal year 2026 (through the March 2026 quarter), Alibaba Cloud AI-related product revenue was 8.971B RMB, eleven straight quarters of triple-digit growth, roughly 30% of external cloud revenue. Qwen is open and free, with token consumption eventually pooling into cloud revenue — as Joe Tsai put it publicly: pure model companies struggle to make money from open source alone; full-stack companies recover the value at the cloud and application layer.
| Dimension | Call |
|---|---|
| Founder-product fit | N/A: a product inside a giant, not a founder project; the decision chain is Alibaba Cloud's AI commercialization strategy |
| Product insight | Upper-middle: splitting AI into high-frequency actions (IME, meeting notes, selection) is an "entry-point" mindset that understands retention better than a lone chat page |
| Execution quality | High: the Qwen family is top-tier among open models, with complete engineering and infrastructure |
| Timing | Stable: the general-assistant landscape is settled (Doubao/Qwen duopoly); the web door is legacy retention, not a growth battlefield |
A channel entry for a stable head product — real data, and genuinely flat.
At 44.53M visits with +0.56% MoM, this is what a mature product's web door should look like: volume without growth. It is not the entry point for growth stories; it is the entry for absorbing PC office scenarios and parking AI into the desktop workflow. Its existence itself shows Qwen is running a combined play: chat app + web + desktop tools.
Read: for the reader, the value here is not growth but structure — multiple doors for one model is not waste, it's testing mindshare. The app acquires, the web retains in office, the IME/selection tools are high-frequency touchpoints. If you have an AI product, copy the split: which entry owns growth and which owns retention — don't expect one door to do both.
The real relationship to track is with the app: if web visits stay flat while the app keeps growing, mindshare lives entirely on mobile and the web door is just a compliance/SEO facade; if web starts climbing, the office scenario is gaining.
① Whether web visits leave the flatline within three months — new features like the office assistant and IME, if they create increment, will show up here first ② Whether Alibaba keeps disclosing the 295M cross-platform MAU (March figure) next quarter and its growth — the merged scope is more meaningful than any single door ③ Whether the Qwen app's MoM (+7.46%) stays positive — mobile is the main battlefield; its movement is what gives the web door a linked story
Product logic: copy the division of roles across multiple doors of one brand — growth door (app/new-user), retention door (web/office), high-frequency touchpoints (IME/selection), all strung on the same model base. Even a small team can build the high-frequency small action as a lightweight client first, which beats building a second chat page.
Pricing structure: none independent. The free strategy is subsidized by the parent ecosystem and not directly transferable to standalone products.
Worth watching. The data is real and the landscape is settled; this is not a growth story but the "office retention" piece of an ecosystem puzzle. Watch not whether it grows by itself but whether its combination with the app keeps pushing the cross-platform MAU higher — and whether the new IME and meeting-note tools actually add desktop-side increment.
05
Use these searches when the official site is missing or the current link is only a lead.