AI is moving from answering questions about universities to acting on them. Agents read the small print, so yours had better be right.

In February, EAB reported that 46% of American high-school students had used AI tools in their college search, up from 26% the previous spring. Nearly one in five had crossed a college off their list because of what the AI told them. I wrote about that invisible shortlist in October.

The next step is already arriving. AI is moving from answering questions to doing things. OpenAI’s Atlas and Perplexity’s Comet are browsers built around agents that can open pages, fill in forms and complete tasks for the person using them, and Amazon has taken Perplexity to court to stop its agent shopping on users’ behalf. The retail trade press has been calling 2026 the year of zero-click buying.

Universities should assume the same thing will happen to the way students compare courses. Some of the people reading your website won’t be people at all.

What an agent sees

A prospective student browsing your site responds to photography, video and tone. An agent asked to “compare business degrees at these four universities on cost, entry requirements and graduate jobs” responds to none of that. It goes looking for facts, and it works with whatever it can find and parse.

That changes what good looks like. Fees buried in a PDF, entry requirements that differ between the course page and the prospectus, outcomes data that’s two years old or missing: a human might persist through all of that. An agent will either get it wrong or move on to an institution whose information is clearer. Either way, you lose the comparison without ever knowing it happened.

For a decade, university websites have been made more emotional and more visual, and rightly so. The facts underneath them now matter just as much, because they’re what a machine will quote on your behalf.

The most underrated reputation asset of the next three years is boring, accurate, consistent information.

An afternoon audit any marketing director can run

Ask three AI assistants to compare your flagship course with two competitors. Ask about fees, entry requirements, placements and graduate outcomes. Check every number they quote against your own site and note where they’re wrong.

Trace each error back to its source. Most will come from outdated pages, conflicting figures across different parts of the site, or information that only exists in PDFs.

Make one page the source of truth for each course. Put the facts in plain text, in the same format for every course, and keep them current. Retire or redirect the pages that contradict them.

Publish your outcomes evidence. If you don’t say what your graduates go on to do, an agent will either fill the gap from elsewhere or leave it blank. Neither helps you.

The opportunity

None of this needs new technology or a big budget. It needs someone to own the accuracy of the university’s public facts as seriously as the brand team owns its look and feel. The universities that do that work now will be the ones agents describe correctly, and in a market where more and more choices start with a machine, being described correctly is a competitive advantage.

I’ll be exploring how admissions teams are preparing for this on Season 2 of the podcast. If you’ve run this audit on your own institution, I’d be very interested to hear what it turned up.

Jonathan Sabarre FCIM MBA is Principal Consultant at QS Quacquarelli Symonds, a keynote speaker and host of Marketing Greats in Higher Education. Get the weekly email or invite him to speak.