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
纳米AI搜索 turns “find links” into handing over a video, report, or deck. Monthly visits: 180 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.
Search is climbing from links to deliverables, and the wider the product, the shallower each line. Don't build another everything-search. Lock “one sentence, finished file” to an industry weekly or a selling video, and make quality good enough to use as-is.
A public record shows 180.29M monthly visits and -8.84% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
① Whether web visits stop falling next month — a second consecutive MoM decline means the; rush is over and the base has settled; ② Whether member revenue keeps being disclosed and growing — the monetization proof that; separates it from other AI search products; ③ Whether app MAU crosses 10M — whet…
Worth dissecting. A public record shows 180.29M monthly visits and -8.84% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
the generic structure of "one-sentence deliverable" is copyable — decompose the task, assign one clearly-role-defined agent per step, run in parallel, reassemble. This "hand it to a team of specialists" frame transfers to any high-frequency "input a need, output an artifact" scenario. keep high-frequency basics free (search/Q&A), and concentrate payment and credits on heavy, high-cost output (per-render video credits). The "free high-frequency + metered heavy-asset" split tracks cost structure better than a one-size membership.
Free for high-frequency basics, paid for heavy-cost output: ① Whether web visits stop falling next month — a second consecutive MoM decline means the; rush is over and the base has settled; ② Whether member revenue keeps being disclosed and growing — the monetization proof that; separates it from other AI search products; ③ Whether app MAU crosses 10M — whet…
纳米AI搜索 turns “find links” into handing over a video, report, or deck. Monthly visits: 180 million.
A public record shows 180.29M monthly visits and -8.84% 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: “纳米AI搜索 turns “find links” into handing over a video, report, or deck. Monthly visits: 180 million.”. 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.
Search upgraded from "find a link" to "give a command, get a deliverable": one sentence in, a video, report, or deck out — powered underneath by 16 models and a swarm of collaborating agents.
Qihoo 360. The product descends from 360 AI Search, launched February 2024, renamed Nano Search, then Nano AI. It is led by Liang Zhihui, a 360 VP from the Safe Guard and browser-assistant teams. This is the flagship C-end product of CEO Zhou Hongyi's "All-in-AI" push, sitting alongside 360 AI Office, the 360 Zhinao model, and security agents.
Read: a security company using free search as a traffic hook and memberships plus credits as the profit layer is a more traditional internet business than most AI startups.
What it explicitly fails at: the mobile app. Web traffic is #2 in China; the app does not crack the domestic top ten by MAU.
"Research then produce" used to be a manual pipeline: open Baidu or Google, click through links one by one, copy-paste into a document, format it into a report or deck. A decent industry report took a skilled person half a day to a day. Nano AI sells compressing that pipeline into one sentence.
More precisely, it replaces the three-in-one stack of general search + Q&A + content creation tools. To be honest: whether it truly replaces anything depends on output quality, and a published hands-on test (Sina Tech, July 2026) concludes that features pile up but each line is only so-so — the shared weakness of every search product climbing up to deliverables.
Free for high-frequency basics, paid for heavy-cost output:
Public reports citing 360's own disclosures put H1 2025 AI value-added services revenue at RMB 905M (+53.65% YoY), over 1.2M members, and ~94% gross margin on that line. Read: the numbers come through broker and stock-talk channels, so discount them — but the membership-plus-credits structure itself is verifiable and closely mirrors the tiered subscription Doubao later shipped.
| Dimension | Call |
|---|---|
| Founder-product fit | 360 owns search and PC distribution; the lead is a veteran of its Safe Guard/browser teams. Real fit, but it is a company product, not a founder's |
| Product insight | "From answer to deliverable" is the right direction; the swarm plus model aggregation is a differentiator, but no single feature line runs deep |
| Execution quality | Zhinao 72B + swarm engine + MCP is real engineering; the comic-drama pipeline's "30–60 min per episode, 90% success" is an auditable metric, not a slide |
| Timing | The search layer is already carved up. Free plus channel distribution bought #2, but the failed app suggests users see a tool, not an assistant |
This is the reference point for the real shape of Chinese AI search. The #2 web ranking is not a bought chart — 222M (April) and 180M (this pool) are third-party traffic numbers, more credible than any marketing copy. The -8.84% MoM is an ordinary pullback, not an anomalous spike, which means the growth push is over and the product has settled into its steady tier.
Read: search is a traffic business, and 360 holds #2 through legacy browser/PC channel distribution plus a free tier. Startups should not enter this lane. The transferable part is the product shape — "model aggregation + multi-agent division of labor + MCP tooling" is an assembly that works in any vertical.
Two real weaknesses: an 8.91M app MAU against #2 web traffic shows users treat it as a use-and-leave tool with no retention scenario; and every feature line is shallow — "do everything" easily becomes "do nothing well." Still, RMB 905M in AI services revenue with high margins proves the "free for traffic, paid for profit" model works in Chinese consumer AI.
① Whether web visits stop falling next month — a second consecutive MoM decline means the rush is over and the base has settled ② Whether member revenue keeps being disclosed and growing — the monetization proof that separates it from other AI search products ③ Whether app MAU crosses 10M — whether the web-#2 / app-out-of-top-10 gap ever closes
Product logic: the generic structure of "one-sentence deliverable" is copyable — decompose the task, assign one clearly-role-defined agent per step, run in parallel, reassemble. This "hand it to a team of specialists" frame transfers to any high-frequency "input a need, output an artifact" scenario.
Pricing structure: keep high-frequency basics free (search/Q&A), and concentrate payment and credits on heavy, high-cost output (per-render video credits). The "free high-frequency + metered heavy-asset" split tracks cost structure better than a one-size membership.
Worth watching. The #2 position in Chinese AI search is real, the traffic data is credible, and the membership model is proven — but the search market itself is locked up. Its sample value is the multi-agent deliverable product shape, not search per se. Write it down and check the three metrics above in three months.
05
Use these searches when the official site is missing or the current link is only a lead.