Why AI search decides education outcomes
46%
of students use AI in their college search
EAB surveyed more than 5,000 US high schoolers in late 2025: 46% now use AI tools like ChatGPT during the college search, up from 26% in spring 2025.
18%
removed a school after an AI answer
In the same EAB study, nearly one in five students dropped a college from consideration based on information an AI search surfaced. A wrong or missing answer costs an inquiry before a human sees it.
92%
of students now use generative AI
HEPI's 2025 survey of 1,041 UK undergraduates found 92% use AI in some form, up from 66% a year earlier. The audience edtech and schools market to is AI-native.
30 of 71
citations went to one directory in our pull
research.com was the single most-cited source across the four 'best LMS / online learning' queries we measured on ChatGPT, 30 of 71 citations (Prefer, Sept 2026).
The buyer prompts that decide education
| Prompt family | An example buyer asks | What wins the citation |
|---|---|---|
| CategorySoftware buyer wants a shortlist | best learning management system for higher ed | A current, structured category page plus a complete listing on the directories AI reads |
| ComparisonBuyer is down to a few platforms | Canvas vs Blackboard vs Moodle | A fair, specific comparison page backed by third-party review corroboration |
| Program discoveryA prospective student is building a shortlist | best online MBA programs | Program pages with outcomes, accreditation and cost that ranking sites cite |
| ComplianceAn admin is gating on data rules | is [platform] FERPA and COPPA compliant | A plain, quotable compliance page; edtech deals stall without it |
| IntegrationBuyer is checking fit with their stack | LMS that integrates with Google Classroom | A documented integrations page in plain language |
| PricingBuyer is budgeting before a call | how much does an LMS cost for a school district | An honest pricing explainer with ranges and what drives cost |
Example prompts are illustrative of each family; run your own category, rivals and personas to build the real set.
Which engines matter for education, and why
- Primary
ChatGPTStudents and educators are among the heaviest ChatGPT users, and our education pull was run on ChatGPT and returned rich directory and vendor data. It is the default first stop for both buyers. - Secondary
Google AI OverviewsCollege-search and 'best LMS' queries frequently trigger AI Overviews, and institutions live and die in Google. Strong SEO overlaps here more than on the chat engines. - Secondary
GeminiGoogle Workspace for Education and Classroom run deep in K-12 and higher ed, so Gemini rides that ecosystem for education questions. - Minor
PerplexityUsed by researchers and graduate students for cited answers. Smaller reach for this buyer, but it rewards clean, quotable program and product pages.
Who AI reads for education answers
- research.comReview siteThe most-cited source in our education software pull. A directory whose 'best LMS' lists AI quotes for category prompts, so placement and accuracy there matter.
capterra.comReview siteA Gartner-owned buyer directory. Profile completeness and category placement decide whether you are named for software prompts.
en.wikipedia.orgReferenceCarries category and vendor entities. A clear, well-sourced page is how a model learns what your platform is and where it fits.
reddit.comCommunityTeacher and admin subreddits trade real recommendations, and Reddit was the #1 source ChatGPT cited across the B2B queries in our Reddit study.- .edu and ranking pagesInstitution / RankingProgram and ranking pages (SNHU, WGU, US News, Niche) are the college-search cluster AI leans on for program-discovery prompts.
- softwareadvice.comReview siteAnother Gartner-owned directory that surfaces for software category prompts alongside Capterra.
- Vendor-owned sitesCompetitor-ownedInstructure, Moodle and D2L win the entity and 'X vs Y' prompts from their own docs and comparison pages, so publish yours.
These source patterns trace to our four-engine AI Citation Study; re-check them as the category moves.
education-specific moves
- Serve both buyers with the right pageEdtech vendors need a category and comparison page; schools need program pages with outcomes, cost and accreditation. Build the one your buyer actually asks AI for, not a generic overview.
- Answer the compliance question in plain textFERPA, COPPA and accessibility gate education deals. A clear, quotable compliance page is the difference between being shortlisted and skipped, and most brands hide it in a PDF.
- Fix your directory and Wikipedia recordresearch.com, Capterra and Wikipedia decide the software shortlist. Complete your profiles, get the category right, and keep the entity page accurate.
- Get into the ranking and roundup sourcesFor programs that is US News, Niche and .edu roundups; for software the 'best LMS' listicles. One accurate placement enters answers your own site cannot win alone.
Education splits into two buyers who ask AI different questions, and the sections above map both: the software prompts an edtech vendor competes on, the program prompts a school competes on, which engines answer them, and who AI reads to decide. Prefer (our product) tracks both kinds of prompt on ChatGPT, Gemini, Perplexity, Google AI Overviews and AI Mode, and shows which directories, ranking pages and threads each engine cites. The honest headline is that most of the deciding sources are not your own site, so the work happens across directories, ranking pages, compliance docs and community threads, per prompt family. The plays follow, in the order a team starting from low visibility should run them.
The plays that win education answers#
Publish the page your buyer asks AI for
Give each of your two buyers the page AI needs to cite: a current category and comparison page for edtech software, and program pages with outcomes, cost and accreditation for institutions.
Education has two buyers who ask AI different questions. Software buyers ask 'best LMS' and read directories; prospective students ask 'best online [program]' and read .edu and ranking pages. EAB found 46% of students now use AI in the college search and 18% have removed a school based on an AI answer, so the page that answers your buyer's actual prompt is what keeps you on the list. (EAB college-search survey, checked 2026-09-05)
- Decide which buyer this page serves, a software category or a program, and write for that prompt.
- Lead with a self-contained answer: what the platform or program is, who it fits, and the one number that matters (price range, outcomes, accreditation).
- For programs, state outcomes, cost and accreditation plainly; for software, state the category, the core features and the main limit.
- Add FAQPage schema to the questions buyers ask so the answer is machine-readable.
Done when: Each buyer you serve has a current page that opens with a liftable answer and carries valid schema.
Verify it worked: Ask your buyer's real prompt in ChatGPT and Google AI Overviews and check whether your page, or only a directory, is cited.
Common failure mode: One generic 'solutions' page for every audience. A model cannot lift a specific answer from it, so it quotes a directory or a rival that published a focused page.
Answer the compliance question in plain text
Publish a FERPA, COPPA and accessibility page that states your posture in crawlable sentences, because compliance is a citation gate unique to education.
District and institution buyers gate on data and accessibility rules, and they ask AI 'is [platform] FERPA compliant' directly. If the answer lives in a gated PDF or behind a sales call, a model has nothing to quote and names a competitor it can read.
- Write a compliance page that answers FERPA, COPPA, GDPR and accessibility (WCAG, VPAT) in plain sentences.
- State what data you collect, where it lives, and who can access it, in text a crawler can read.
- Link your VPAT and sub-processor list as real pages, not only downloadable files.
- Add the compliance questions to a schema-marked FAQ.
Done when: Your FERPA, COPPA and accessibility posture is stated in crawlable text with a schema-marked FAQ.
Verify it worked: Ask 'is [your platform] FERPA compliant' in a grounded engine and check whether it can state your posture.
Common failure mode: Compliance locked in a downloadable PDF. Crawlers skip most PDFs, so the model treats the question as unanswered and moves on.
Fix your directory and review record
Complete your research.com, Capterra and Software Advice profiles with the right category, because those directories are the sources AI quotes first for education software.
Buyer directories are the pages engines read first for 'best LMS' prompts. In our education pull, research.com alone took 30 of 71 citations across the four queries we measured. Your category placement and profile completeness there feed straight into the shortlist. (Prefer AI Citation Study, checked 2026-09-05)
- Claim and complete your research.com, Capterra and Software Advice profiles with an accurate description.
- Get the category right: check which category your rivals sit in and match the one buyers ask about.
- Run a review drive to raise volume and recency so the model has current material to quote.
- Read recent reviews for one recurring criticism and fix it at the source.
Done when: Your directory profiles are complete, correctly categorized and gathering fresh reviews.
Verify it worked: Ask 'best [your category] software' and check whether the directories cited now list you accurately.
Common failure mode: A half-finished profile in the wrong category. Engines read the directory literally, so a mis-tagged listing is worse than none.
Earn placements in the ranking and roundup sources
Get accurately listed in the ranking sites and roundups AI cites, US News and Niche for programs, 'best [category]' listicles for software, because those enter answers your own site cannot win alone.
For 'best online [program]' and 'best LMS' prompts, AI leans on third-party rankings and roundups more than any single school or vendor site. An accurate placement in US News, Niche or a category listicle puts you in answers your homepage cannot reach. (Prefer Reddit study, checked 2026-08-19)
- Run your category or program prompts and note which ranking sites and roundups get cited.
- For programs, make sure your US News, Niche and accreditation listings are complete and accurate.
- For software, pitch the trade publications behind the 'best [category]' roundups with a specific reason to include you.
- Take part honestly in the teacher and admin subreddits where educators trade recommendations.
Done when: You are accurately listed in the top ranking sites and roundups AI cites for your category or program.
Verify it worked: Ask your category or program prompt in ChatGPT and Google and check whether the sources cited now include you.
Common failure mode: Planted or spammy self-promotion in community threads. It gets removed, and a recommendation that reads as planted persuades neither a reader nor the model.
How this fits your existing SEO and enrollment work#
None of this replaces SEO or the enrollment funnel you already run. The same category, program and compliance pages that help you rank also feed AI answers, so a crawlable, authoritative site is a head start. What changes is the target: SEO earns a ranking for a keyword, AEO earns a citation inside the answer, and the citation rewards specificity, plain structure and third-party corroboration over keyword coverage. Because these education prompts have almost no Google volume, judge the work by whether the answer names you, not by rank.
For the full method behind these plays, start with what AEO is, compare your model against the AEO for B2B SaaS playbook since edtech software behaves like it, or browse every AEO-by-business-model playbook.
A worked example
Holds clear category-leader positioning on the software directories AI reads for 'best LMS', so it is named before smaller rivals.
Open-source, with a strong Wikipedia and entity footprint that wins definitional and 'X vs Y' prompts.
A course-platform brand with heavy earned coverage, cited across both program and platform answers.
Southern New Hampshire University
An institution that markets online programs and surfaces in the college-search cluster AI cites for program prompts.
Named brands are public, illustrative examples of the category, not customers or endorsements.
Sources
People also ask
- How do edtech companies get cited by ChatGPT?
- How do colleges show up in AI college search?
- Which sources does AI use to recommend education software?
- Does FERPA compliance affect AI recommendations?
- How is AEO different from SEO for edtech?