BOSS Zhipin is probably done for.
The day before yesterday, I read a rant about BOSS Zhipin. The author used AI to help screen resumes and got flagged by the platform's risk control system. The most absurd part of this story is that BOSS itself has a paid feature called "Talent Bomb"—which sounds like "border conflicts for workers"—but its function is to send greetings to a bunch of people at once.
The same capability: when users implement batch screening through an AI browser, it's called cheating; when the platform sells it as a feature, it's called product capability. When users save time with AI tools, it's "abnormal risk"; when the platform packages the same thing as a paid feature, it's called efficiency improvement.
So the question is: does the platform prohibit efficiency, or does it prohibit you from getting efficiency for free?
The interesting part of this story isn't that some user got banned—after all, internet platforms flagging users is hardly news. What's truly worth discussing is that it perfectly illustrates the business logic of recruitment platforms.
Platform economies appear to facilitate transactions, matching companies and job seekers within the platform. But in reality, their business model is advertising space rental, or even an "exposure tax."
Companies pay for access to more talent; job seekers pay for more resume exposure. But the platform isn't directly responsible for "whether the match succeeds"—it cares more about whether both sides continuously generate exposure, clicks, and interactions during this process.
Letting companies almost hire the right person, letting job seekers almost find the right job—both sides maintain an ambiguous boundary, like an unattainable lover. So both sides have to keep paying for services and exposure. This is the most stable state for recruitment platforms.
So, it's exactly right that companies struggle to hire.
What the platform sells you isn't successful matching, but "keeping you recruiting forever." Resume packages, exposure, refreshes, top placements, bulk outreach, job promotions—these things don't solve whether the "matching is precise," but rather "can you see more people," "can you be seen by more people," "can you send more messages."
What recruitment platforms sell to companies isn't "the right people," but "the opportunity to contact more people."
But what companies actually need is less noise, less ineffective communication. Not automatically refreshing jobs daily, but quickly knowing who is really suitable and who isn't worth talking to at all.
On the other side, for job seekers wasting away their days in the platform, of course it's hard to find work.
When job seekers open the recruitment platform, they see many opportunities, many positions, many companies, many recommendations. The system tells you which positions suit you, which are trending, and which HR just became active. The entire interface is as bustling as a village market; through the screen you can hear countless HR vendors shouting, as if just a little more effort will land you a job.
But the real experience is sending many applications, greeting many people, and receiving a pile of "read but no reply" or even "not read, no reply." You see positions permanently "dominating the charts" in recommendations, hanging there year after year like insatiable gluttons—however many resumes you throw in, you get no response. You see salary ranges spanning two orders of magnitude, where anything above the base salary turns out to be dreams and future promises when you ask.
You revise your resume, refresh your resume, optimize keywords, adjust job intentions, and still don't know where the problem lies. This uncertainty creates anxiety that traps job seekers in the platform as active users, becoming the platform's assets.
Anxiety has no benefit for job seekers, but it increases platform activity and stickiness, and sells a bunch of refresh tools, ranking boosters, resume editing, and mock interview services.
Two problems, one business.
Talent can't be recruited, jobs can't be found—two groups trapped in the same system staring at each other. Companies curse job seekers as unreliable; job seekers curse companies as blind. Both sides think it's the other's fault; nobody thinks to blame the platform that matched them.
This isn't just BOSS Zhipin's problem; it's the ultimate form of all middleman businesses. Information asymmetry is the source of profit, inefficiency is the moat, anxiety is the business model.
What platforms do isn't true transaction matching—they don't actually like true matching, because successful matching means the transaction ends. They like the state of "matching in progress": companies always recruiting, job seekers always searching, both sides always just one breath away from each other.
This breath is the platform's ventilator.
So don't expect the platform to save you with AI. Its AI will only be used to keep you hanging on that breath a little longer, recommending more precise "almost theres," manufacturing more realistic "almost ready" illusions. Companies' recruiting budgets are spent, job seekers' patience is exhausted, but the platform's financial reports look prettier.
This isn't a bug, it's a feature.