The launch lands during the post exam college preference week, when Chinese families pay thousands of yuan for advice that often amounts to 'hot majors' and 'insider admissions data' scams.
The 12.9 million Chinese students who sat this year's Gaokao, the national college entrance exam, received a new piece of free advice on June 10, and the timing was deliberate. Alibaba's Qianwen AI advisor went live as scores were released, offering every test-taker a tiered list of college recommendations for the form they must fill out within days: six "high-potential" schools to reach for, eight "stable" choices, and sixteen "safety-net" options. The product is free, public, and tied to one of China's largest general-purpose consumer AI brands, which is why the launch is being read less as a technology release and more as an entry into a market that has preyed on families for years.
The form in question is 志愿填报, the post-Gaokao college-preference form that the Shuzi Lichang feature describes as "the second-most stressful time for a kid, behind only the exam itself." Once scores are in, students typically have a few days to rank up to five colleges and three to five majors each, working from opaque enrollment quotas and government-set cutoff scores that do not guarantee admission. The matching rules are tangled enough that the order of the rankings, and the majors paired with them, can swing outcomes. A 2021 working paper by Ruixue Jia and co-authors found "great room for improvement in [student-college matching quality]" in the current system, supplying an empirical baseline for any intervention that claims to help.
That is the gap Qianwen is selling into, and the company is not the first to try. The Shuzi Lichang piece, translated and contextualized by Jeffrey Ding's ChinAI newsletter, documents a long-standing advice ecosystem that is expensive and often low-quality. A relative in the article paid 10,000 yuan (roughly $1,400) for a consultant whose final output was a list of "hot" majors. The wider market is full of operators selling "insider admissions data" and the services of supposed "exam-setting professors," a category of fraud that has grown up precisely because the official information is thin and the stakes are high. The original Chinese feature frames Qianwen's move as turning "college application guidance from a business based on information asymmetry into an inclusive public service," an analogy Ding draws to the earlier Zebra English AI tutoring case, in which a general-purpose consumer AI undercut a paid test-prep industry.
Whether the substitution actually improves matching is the question the launch does not answer. The Qianwen output is structured around the 冲/稳/保 convention familiar from Chinese college-prep materials, a tiered risk profile that human counselors have long produced by hand. The free, mass-reach claim depends on Qianwen's brand recognition as a general-purpose assistant, per the original author, and Ding flags the back half of the Shuzi Lichang piece as "press-release-y." The 12.9 million figure is the size of the 2026 test-taker cohort that Qianwen has made the service free to, not an audited count of students who actually used the tool, and it should be read as a reach number, not an adoption number.
The most cited concern under the WeChat post, with 26 likes at the time of Ding's translation, is homogenization: if every student in the same academic tier receives the same AI recommendation, the tool could push the same colleges and majors to the same applicants and degrade the matching it is meant to improve. That tension is real and is sourced to a single commenter's worry, not a study, but it sits directly on top of the matching problem the Jia and Li paper documents. The constructive question is not whether Alibaba's AI is impressive; it is whether a single platform's model can replace the existing ecosystem of brokers, teachers, and parents without collapsing into a different kind of gatekeeper.
The full English translation of the Shuzi Lichang feature is available on Google Docs. What to watch: whether Qianwen publishes any matching-quality outcome data for the 2026 cohort, and whether Chinese education authorities treat the free tool as a complement to official admissions channels or as a competitor that needs new disclosure rules.