The front door moved, and nobody sent a change of address card
Picture a high school senior at eleven at night, laptop open, trying to narrow forty schools down to eight. She is not opening ten browser tabs. She is not comparing viewbooks. She opens ChatGPT and types something like, "what are the best affordable nursing programs for someone who wants to work while they study." She gets one answer. A short list. Maybe four names.
That is the moment your entire enrollment funnel now runs through, and almost no institution has built anything for it. Admissions offices spent two decades perfecting the glossy brochure, the campus tour, the viewbook photography, the paid search campaign aimed at ranking on page one. All of that infrastructure was built for a world where a student clicked through ten results and formed her own opinion. That world is thinning out fast. A growing share of students now get a single synthesized answer and never see the ten results at all.
This is not a hypothetical shift happening at some other, more digitally forward institution. It is happening inside your applicant pool right now, tonight, to a student you have never heard of, about a program you are proud of.
What happens when you ask AI about your own school
Try it. Open ChatGPT and ask it to recommend Christian colleges in the South with a strong music business program. That is a real test one enrollment VP ran on his own institution, a Christian university in the South with music business as its flagship major. The AI didn't mention his school. Not on the first try, and not after he clarified the major.
That is not a broken system. That is the system working exactly as designed, and simply not finding enough clear, extractable evidence that this school belongs in the answer. The model is not being unfair. It is being literal. It can only recommend what it can confidently find, understand, and trust. If your program pages describe outcomes in admissions language instead of the language a 17-year-old actually uses, the model has nothing clean to work with, and it moves on to a competitor who gave it something usable.
The 18% you never hear from
Here is the part that should genuinely unsettle an enrollment team. Eighteen percent of students say they have already removed a college from their list because of something an AI chatbot told them. Not because they toured campus and didn't like the dining hall. Not because financial aid fell through. Because a language model gave them a confident, synthesized answer, and that answer didn't include your school, or worse, included it with something inaccurate attached.
Nobody files a complaint about this. No application gets marked "declined, reason: AI told me not to apply." It shows up nowhere in your CRM. It just shows up, quietly, as an inquiry that never arrives, an application that never gets submitted, a shrinking top of funnel that your team spends the next planning cycle trying to explain with the usual suspects: demographics, competition, a weak digital campaign. Meanwhile the actual cause was sitting in an AI conversation that happened on a Tuesday night that nobody in your building will ever see.
This tracks with the broader shift in how this generation actually behaves. Roughly two out of three US teens already use AI chatbots, per Pew Research, and for them this is not a novelty tool. It is simply how you get an answer to something you're not sure about, the same instinct that used to send a previous generation to a guidance counselor's office.
This isn't SEO with a new coat of paint
The instinct is to hand this to whoever already handles search engine optimization and ask them to "do the AI version." That undersells how different the mechanics are. Traditional SEO optimizes to rank in a list of ten blue links a human will scan. AEO optimizes to be the specific sentence a model lifts out, restates, and attaches your name to, inside an answer the student never has to click through to see.
That changes what good content looks like. A model rewards atomic, self-contained facts over sweeping brand narrative. It rewards a direct answer positioned at the top of a page, with supporting detail underneath, over three paragraphs of scene-setting before you get to the point. It rewards consistency: your program's tuition, format, and admission requirements need to say the same thing on your website, your CRM-facing landing pages, and any directory or aggregator that lists you, because a model that catches you contradicting yourself has every reason to trust you less.
None of this comes with a published rulebook. No AI company hands out a ranking formula the way Google eventually leaked pieces of its algorithm over the years. What we have instead is a pattern that keeps repeating across every institution that has actually tested this: clear, specific, current, and consistently structured content gets cited. Vague, aspirational, marketing-department content does not.
What actually gets a program cited
Start with the questions your own admissions counselors already answer every single day on the phone. That call log is the single best keyword research tool you have, and almost nobody mines it. "Can I transfer in with an associate's degree." "Is the program accredited for my state's licensing board." "What does a typical week look like if I'm working full time." Those are the actual questions. Your program page probably answers a more polished, more institutional version of them, worded the way a catalog committee would word it, not the way a nervous 19-year-old actually asks.
Write the direct answer first. Put "yes, you can go part time, here is what that schedule looks like" in the first two sentences under the heading, not buried in paragraph four after a section on the university's founding history. Structure the page around the student's real question as the heading itself, not a generic label like "Program Overview." Keep your facts, your tuition figures, your outcomes data current everywhere they appear, because a model checking your page against a directory listing from two admissions cycles ago will simply trust the fresher, more consistent source instead.
And build content depth around the topics that actually generate volume, rather than spreading thin across every program you offer. This is the same principle behind our own 12 Anchors approach: find the one question that a real, meaningful number of prospective students are actually asking about a given program, build something substantial that genuinely answers it, and let that one strong piece of content do the work of ten shallow ones.
The uncomfortable part
Here is the oblique truth in all of this. For years, "digital marketing" for a college meant making the institution look as good as possible: the best photography, the most polished copy, the most persuasive framing. That instinct is now working against you in a specific, measurable way. A page engineered to impress a human reader with tone and narrative is often the exact page a model struggles to extract a clean answer from. The schools winning this moment are not necessarily the ones with the biggest marketing budget. They are the ones willing to write plainly, answer directly, and let the substance carry the persuasion instead of the polish.
We've watched this exact problem up close working with institutions like LECOM, where the real question is never "how do we look impressive," it's "does a prospective student, or an AI system standing in for one, get a clear, honest, immediate answer to the specific thing they're actually asking." That is a solvable problem. It just isn't solved by another round of brochure copy.