How Students Use AI to Choose Colleges (And Why It’s Happening Earlier Than You Think)
According to EAB research, 46% of students already use AI in the college search process, and 62% use it to find best-fit schools. Nearly 1 in 5 students have removed a college from consideration based on AI-generated insights.
At the same time, Encoura data shows 43% of students decide by fall of senior year or earlier.
The implication: Students are forming college shortlists influenced by AI before most university marketing begins, and institutions that are not clearly understood by AI systems are often excluded from consideration.
Most universities behave as if the enrollment journey begins when a student starts actively searching.
Junior year. Campus visits. Inquiry forms.
That’s when we show up.
The problem is, by then, the decision is already taking shape.
By the time a student fills out a form, something quiet has already happened. They’ve formed a sense of which schools feel like them. Not a ranked list. A shape. A few places that feel right. A few that don’t.
That shape forms earlier than most institutions plan for. Sophomore year, sometimes sooner. Not during a campus visit. Not during a campaign.
It happens when a student is in their room, on their phone, trying to answer questions they don’t fully know how to ask yet.
They’re not searching for your university. Yet.
They’re trying to understand themselves within the context of this new life stage. With AI.
And once that match is made, a decision has formed. This shift is happening before your recruitment cycle begins, and outside the channels you normally measure today.
Encoura data shows 43% of students decide by fall of senior year or earlier, up from 19% just a few years ago.
The timeline is compressing. Students form preferences earlier. They lock faster. And with AI intermediating, they revisit less.
The Moment No One Measures
Enrollment teams will tell you students are harder to move than they used to be. A more accurate read: they’re not deciding when we meet them. They've already decided. They’re now confirming.
By the time a student engages with an institution, they already have a shortlist based on feelings.
That sorting happens early. Once it sets, it holds.
Institutions are losing visibility in the exact moment students form that emotional shortlist. If a student encounters your school early, you anchor the decision. If they encounter you later, you don’t compete for preference. You compete for fallback.
You become the safety school.
How That Exploration Now Works
This isn’t new behavior. Students have always started forming impressions early. What’s changing is how that early thinking now evolves within AI chatbots. Students don’t move through a funnel anymore. They move through a sequence.
Chat histories builds context in ways that Google Search never did. Each question builds on the last. Context carries forward. Options narrow.
Students refine. AI interprets. The list narrows. They're not search for institutions yet. They’re building a narrative about themselves and where they belong in this world.
According to EAB, 46% of students already use AI in this process. Among them, 1-in-5 have removed a school based on what AI returned.
Institutions are no longer competing to be discovered.
They’re competing to be included.
Where Institutions Break Down
Strong brands will still show up.
If you’re MIT, you’re not fighting for inclusion. You’re assumed. That doesn’t change.
But most institutions don’t live in that category. And even for those that do, what shows up alongside them, and how they’re framed is still very much in play.
Because this isn’t just about reputation. It’s about whether an institution can be clearly understood, consistently described, and easily matched to the kinds of questions students are asking.
That sounds abstract until you look at where it breaks.
Students are asking questions about themselves.
Most universities describe themselves well. That’s not the problem.
The problem is alignment: Students ask questions about themselves. Universities answer questions about the institution. Often these don’t match.
Creative reflects this first.
Mission statements and values are great, but they don’t resolve the student’s uncertainty.
A lot of university content explains what the institution is. Mission statements, values, differentiators. They don't explain who the student is within the institution. (“Can I be ambitious here without losing myself?” or “Is this a place where I’ll feel supported or judged?”)
The job isn’t to translate mission into marketing language. The job is to connect what the institution actually is to what the student is already worried about.
That’s a different brief.
UX reflects the same gap.
Most sites are organized around the institution: departments, programs, requirements. A student is arriving with a different question: “Is this a place for someone like me?”
Institutions answer “what we offer” but prospective students are looking for "where do I fit?"
That mismatch isn’t just a content issue. Pages need to connect coherently and semantic meaning needs to carry across pages, otherwise signals fragment and context resets. This is where AI systems struggle to include you.
And when a student does arrive from an AI-generated recommendation, is your site designed to continue the conversation? Behavioral signals matter.
No clear next question + no guided path + no interaction that helps them refine fit = an experience that breaks at the exact moment interest forms.
Tech compounds it.
In some cases, your site is simply invisible to AI.
Security layers designed by IT teams to block bots for the past decade now often block AI crawlers as well. WAF rules, IP filtering, and bot mitigation systems can prevent platforms like ChatGPT or Claude from accessing your website at all. In these cases the system infers your content based on outside sources.
Even if not blocked, many AI systems struggle with your site structure.
If you stripped your site down to its underlying components, program descriptions, outcomes, testimonials, schema, could a system accurately summarize your institution from that alone?
For most schools, the answer is no.
The "meaning" lives in enrollment brochures. In one-off copy. In outdated PDFs. In pages that don’t connect cleanly to each other. In concepts that don't reinforce each other. These fragments blur the lines of meaning, and decrease your chances of being included.
Most institutions assume they’re being considered. Many aren’t even being retrieved.
The Game Has Changed
It’s not enough to have strong content. Institutions need to "mean" something. To resolve into something that can be understood quickly, reinforced by what others say about them, and clearly connected to the kinds of questions students are asking in the first place.
It's not enough to have a strong campaign either. You may be reaching prospective students with the perfect marketing message, but arriving after the decision has already been formed. Instead you'll be competing for the "safety school" slot than than the school of preference. A slot that demands higher acquisition costs and results in a lower ROI.
When you see your YoY traffic numbers decline, and your enrollment goals further out of reach. Ask yourself, are you failing because your campaigns is weak? Or are failing you because you're hard to interpret?