Applicants acted on entry requirements
the university had already changed.

01What happens.

A university publishes across faculties, campuses and decades. Prior-cycle prospectuses, course pages and PDFs stay reachable, and AI answers applicants from whatever it finds. Which answer an applicant receives depends on the tool they use.

02Why it matters.

Your applicants ask AI before they ask you. In a UCAS survey of 4,485 students, 73% of those who used AI received incorrect information, including wrong grade requirements. An applicant decides on the answer they receive, and when it is wrong it comes back to you after results day as a complaint, an appeal or a question about the accuracy of what you publish.

The applicant who is put off never contacts you, so the complaints you receive understate the loss, and the applicants you do hear from are the ones the university can least afford to have misled.

03The case.

This illustrative scenario draws on patterns observed across university estates. It also reflects an observed instance in which applicants received different course information from different AI tools; no university is identified.

A university revises a popular course for the coming cycle: the entry requirement changes by a grade, the fee rises and a related course is withdrawn. Admissions updates the central course page and the online prospectus the same month and publishes the new prospectus PDF on the same page. Admissions sees the update as complete.

The previous cycle’s prospectus PDF stays where it has been for 2 years, on a faculty page and in the university’s document store, and copies sit on school and agent websites that downloaded it, outside the university’s control. The faculty’s own course page is not updated, because faculties and admissions publish separately. The central update is correct, and the wider university estate still carries the old position.

Applicants asking AI about the course receive the previous cycle’s requirements, fee and course list. Some, told the grade requirement is higher than it now is, apply elsewhere. Others, told it is lower, apply and are refused on results day. Some apply for the withdrawn course and are told at the point of offer that it no longer runs. The applicant shown the higher requirement never applies, so the university never knows it lost them.

Which answer an applicant receives depends on the tool they used. One tool serves the current page and another serves the old prospectus, and the applicant has no way of knowing that a difference exists, because each answer arrives with the same certainty and neither mentions a date. Admissions, checking the updated central page, sees the correct position.

The complaints come after results, and with them the questions a university has to answer about the accuracy of the material it publishes to applicants. With the old prospectus still reachable, the following cycle repeats it.

04Exposure.

A university is held to what it publishes to applicants, on whichever page or document it was published and however long ago. Where an old prospectus is still reachable, the requirements it states are still the university’s published position to the machine reading it, and the applicant who relied on it has a question the university has to answer.

The cost is complaints handled, appeals heard, offers reissued and applicants lost to competitors on requirements the university did not set. The larger cost is the applicant’s confidence in what the university says, at the point where the applicant is choosing between universities.

05What AI Readiness changes.

What the university cannot see from the inside is the old prospectus still answering for it, because the central page it checks shows the correct position. AAAnow reads the university’s digital landscape as AI reads it and evidences which prospectuses, course pages and documents AI is drawing on, including copies outside the university’s own sites. With that evidence the university can decide, course by course, which position stands, retire the contradicting material it controls in a form a machine reads as superseded, know where the copies it cannot retire sit, and check the position before the cycle opens.

Your team makes the changes; AAAnow provides the assessment, the evidence, the guidance and the review.

06Related use cases.

07References.

• UCAS survey of 4,485 students, reported by Times Higher Education, 2025. www.timeshighereducation.com/news/most-uk-applicants-used-ai-university-decisions-ucas-polltimeshighereducation.comWe do not enable external links - please copy and paste, you are then certain as to the link being opened