Home
How is college discovery actually changing?
Start with social search, which stopped being novel years ago. As far back as 2022, Google's own leadership acknowledged internal research showing roughly 40% of young adults turned to TikTok or Instagram instead of Search or Maps for everyday discovery, like finding somewhere to eat. Industry surveys since have consistently put Instagram and TikTok ahead of Google for local discovery among 18-to-24-year-olds. Gen Alpha inherits this as the default, not the disruption.
Then there's AI, which crossed into the mainstream of college search specifically — and recently. EAB surveyed more than 5,000 high school students in late 2025 and found 46% now use AI tools such as ChatGPT, Gemini or Perplexity in their college search, nearly double the 26% from just that spring. Among those users, 62% asked AI to find colleges that fit them. A third discovered a school they hadn't considered. And 18% removed a college from their list based on AI-generated results.
The generation behind them is even deeper in. Pew Research Center found in December 2025 that 64% of U.S. teens use AI chatbots, with about three in ten using them daily. For Gen Alpha, asking an AI isn't a research strategy. It's just how you find things out.
Put those together and the conclusion is uncomfortable but hard to dodge: a growing share of your first impressions are being made by systems and creators you don't control, on surfaces you may not even monitor.
What are AEO and GEO, and why do they matter for higher ed?
AEO — answer engine optimization — is the practice of structuring content so answer engines like Google AI Overviews, ChatGPT and Perplexity can extract, summarize and cite it accurately. GEO — generative engine optimization — is the closely related discipline of earning visibility inside generative AI responses. In practice the two overlap heavily, which is why Boxcar refers to AEO/GEO as one capability.
Traditional SEO earned you a ranking. AEO/GEO earns you a citation — or at least an accurate summary. The difference matters because answer engines compress. When a student asks for the best nursing programs near NYC with strong NCLEX pass rates, the AI returns a short synthesized answer naming a handful of institutions. You're either in that answer or you're invisible, and there's no page two.
The stakes are already showing up in enrollment data. When EAB found 18% of students striking colleges from their lists based on AI results, the culprit was often stale or ambiguous institutional content — old tuition figures, buried outcomes, program pages that never directly answer the questions students ask. AI systems fill gaps by inference, and inference is where errors live.
The old rule was that SEO should be invisible to the reader but obvious to Google. The new rule extends it: your content should read naturally to a 17-year-old and parse cleanly for a machine summarizing you to that 17-year-old.
Where does social search fit?
Social search is where the emotional shortlist forms. Students don't search TikTok for "retention rate." They search "day in my life [university]," "dorm tour," "is [college] worth it." What they find is peer content — authentic, unpolished, algorithmically matched to them — and it carries more weight than any viewbook.
Pew's data explains why this layer is unavoidable: roughly nine in ten teens use YouTube and about six in ten use TikTok and Instagram, with a fifth on TikTok or YouTube almost constantly. That doesn't mean every college must be on every platform. It means every college should know what students find when they search these platforms — and should be feeding the ecosystem with platform-native content where their audiences actually are.
The attention economy rewards agile, authentic owned storytelling. Polished 16:9 commercials cropped vertical are neither.
How should colleges structure content for answer engines?
Answer-first structure is teachable, and most of it is unglamorous:
- Direct answers up top. A 40–70 word summary answering the page's core question before the narrative begins.
- Question-format H2s that mirror what students actually ask: "How much does the nursing program cost?" rather than "Investment in Your Future."
- Facts stated plainly, in text. Tuition, deadlines, requirements, outcomes — not buried in PDFs or images AI can't reliably parse.
- FAQ blocks and schema markup (FAQPage, BlogPosting, Organization, Course where relevant) so structure is machine-explicit.
- Freshness and consistency. Answer engines penalize contradiction. If three pages list three tuition figures, the AI picks one — possibly the 2022 one.
- Entity clarity. Name the institution, program, location and credential consistently so systems connect your content to the right entity.
What to do now
Discovery work rewards seeing before planning, so start by looking:
- Search yourself in all three layers. Your top three programs, on Google, on TikTok, in ChatGPT. Screenshot what comes back — that's your real first impression, and it usually surprises people.
- Fix the facts before anything fancier. Getting tuition, deadlines and outcomes correct and consistent everywhere they appear is the cheapest AEO win available.
- Rebuild one program page answer-first and treat it as the template: direct answer block, question headings, FAQ, schema.
- Give someone the listening job. Somebody on your team should know what "day in my life" content exists for your campus — and what conspicuously doesn't.
- Baseline your AI referrals. Traffic from ChatGPT, Perplexity and AI Overviews is trackable, and you'll want a before number when it's time to prove progress.
