How students choose colleges in the age of AI search.

The questions are already written; communities tell you which ones decide enrollment.

The paper argues a community has two distinct uses: a demand signal that reveals the questions worth answering in students' own phrasing, and an alignment instrument that tests whether an institution's content answers them in recognizable terms.

AI search is now a standard, recurring step in the college search — not an edge case.

The anchor study for the thesis: a March 2025 survey run jointly by UPCEA and the SEO agency Search Influence. Being cited in an AI answer measurably raises the odds a student trusts and considers an institution, which places AI visibility upstream of consideration rather than alongside it.

Independent journalism — not a vendor — confirms families are already choosing colleges with ChatGPT.

Because this is reporting rather than a white paper, it is the most credible single reference for a skeptical audience. It quotes enrollment leaders on the shift to "no blue links," and flags that when AI surfaces stale pages, accuracy becomes the institution's responsibility.

When AI answers in place, the click — and the website visit — never happens.

The Chegg figure makes the abstract concrete. The same dynamic that nearly halved its non-subscriber traffic applies to any institution whose story lives only on its own website, reframing AI visibility as a defense against invisibility.

Reddit is the single most-cited domain in AI answers — about twice Wikipedia.

The keystone fact for why Reddit specifically. It comes from Reddit's own Q2 2025 shareholder letter, reported independently, which means the content shaping an institution's AI reputation largely lives on a platform it can participate in openly.

Citation share differs by engine, so AI visibility must be measured platform by platform.

Profound's analysis of 680 million citations is the large-N backbone behind the Reddit claim, and adds an operational lesson: ChatGPT leans on Wikipedia while Perplexity and Google AI Overviews lean heavily on Reddit. "AI" cannot be treated as one monolithic channel.

AI reaches for community content on judgment questions — a structural preference, not a fad.

The questions students bring to AI ("is it worth it?", "are people happy there?") are experience-and-judgment questions that forums answer better than any brochure — which is why the dynamic should persist as engines mature.

There is an official, compliant way to do this — and it is the path we take.

We cite Reddit's sanctioned business tooling deliberately: our approach is transparent, branded participation using approved methods — never scraping, astroturfing, or individual-level profiling. That distinction protects an institution's credibility.

Awareness is near-universal; execution is rare. The early movers compound an advantage.

The readiness poll is the urgency argument. With most institutions only "exploring," the field is open for those that act now — and visibility built early keeps compounding while peers deliberate.

A major, conservative higher-ed firm says outright that AI relies on third-party sites — including Reddit.

Coming from EAB rather than a niche vendor, this validates both halves of the thesis: the steep Gen Z adoption curve and the direct line from third-party and Reddit content to an institution's AI brand authority.

A persona should be a recurring pattern in the data, not a character someone wrote.

The paper argues a segment earns a place in the model only when the same decision behavior recurs across independent sources, making it falsifiable: as behavior shifts, segments are refined or retired. That is the difference between a model and a slide.