Insurance AEO: How Carriers and Brokers Are Losing Quote Volume to AI Search

In April 2026, the National Center for Education Statistics released its updated school choice tracking report, and one number reframed how K-12 marketing teams are thinking about discovery. Across the 13.4 million U.S. households that evaluated a school change in the 2025-2026 cycle — private to public, public to charter, in-district to out-of-district, homeschool to learning pod — 47% used a generative AI assistant at some point in the research process, up from 19% the prior year. ChatGPT alone was named in 38% of those sessions. The discovery surface for K-12 education has moved faster than any operator in the sector expected, and the citation patterns have hardened into a power law that mirrors what we have seen in higher ed and bootcamp discovery.

We spent the last twelve weeks auditing AI citation behavior across roughly 3,000 K-12 queries — schools, tutoring, enrichment, summer camps, learning pods, and microschools — across ChatGPT, Claude, Perplexity, and Gemini. The picture that emerges is striking. Two aggregators (GreatSchools and Niche) dominate the citation layer for schools. A small handful of tutoring brands (Varsity Tutors, Outschool, Wyzant, Sylvan) own the tutoring category. Summer camp discovery has fragmented across regional parenting publications and the ACA accreditation database. And across all three, the independent school, tutor, or camp that wins citation share is the one that has built operator-grade content infrastructure — not the one with the largest marketing budget.

This is the playbook for K-12 operators in 2026: how to read the AI citation patterns that already exist in your category, where to invest to break in, and which surfaces are actually load-bearing for the discovery decisions that drive enrollment.

The GreatSchools Citation Moat

GreatSchools is the single most-cited source in K-12 AI search. Across our 3,000-query audit, GreatSchools appeared in 71% of school discovery answers on ChatGPT, 68% on Perplexity, 54% on Claude, and 79% on Google's AI Overviews. No other source in K-12 even approaches that concentration. Niche.com — the second-place aggregator — sits at roughly 38% across the four assistants, primarily for private school and student-review content.

The dominance is not accidental. GreatSchools has spent two decades building the structural assets that AI models prioritize.

Standardized data coverage at national scale. GreatSchools has stable, normalized data on nearly every K-12 school in the United States — over 130,000 public schools and 30,000 private schools. The data includes standardized test results from state assessments, demographic breakdowns, student-teacher ratios, graduation rates, and college readiness indicators. AI models prefer single-source coverage because it eliminates the need to reconcile contradictory data, and GreatSchools is the only source in K-12 with that kind of horizontal reach.

Extraction-friendly URL and schema structure. Every school on GreatSchools has a stable URL of the form greatschools.org/[state]/[city]/[school-id]-[school-name]. The school profile pages render server-side with structured ratings, demographic tables, and parent review text exposed as HTML rather than JavaScript components. The 1-10 rating scale is consistent across schools, which gives AI models a clean comparison axis they can quote without hedging.

Decade-old SEO equity. GreatSchools has been the default school information source in Google's local school knowledge panel since approximately 2014. Ten years of inbound links, citations from news outlets, and integration with Zillow and Redfin real estate listings have built the kind of entity authority that AI models inherit when they learn the K-12 category. When a model is uncertain which school information source is canonical, it defaults to GreatSchools because the surrounding training data overwhelmingly references it.

For K-12 operators, this has two implications. First, the school's GreatSchools profile is functionally part of the school's marketing surface, whether the school treats it that way or not. AI models quote the GreatSchools description before they quote the school's own website. Schools that have not claimed, audited, and updated their GreatSchools profile are letting an unmanaged third-party page define their brand in AI search. Second, breaking the GreatSchools default requires building competing entity signals — the school's own website needs to expose the same structured data with greater specificity and freshness than the aggregator can match.

The Real Parent Query Behavior

The conventional wisdom about how parents research schools — they ask friends, they tour, they read reviews — is still partly correct, but the AI-assistant layer has been inserted between the initial consideration and the tour. Our interview data, combined with usage analytics from three regional parenting publications that share their AI referral data with operators, shows a consistent five-stage flow:

Stage 1: Category framing. The parent opens ChatGPT and asks a broad question — best Montessori school in Austin, how do I evaluate a charter school, what is the difference between IB and AP. The AI response sets the parent's mental model for the category and names three to five reference points. The schools or methodologies named in this opening response disproportionately anchor the rest of the search.

Stage 2: Local shortlist. The parent narrows to their geography and asks for a list. Best private elementary school Brooklyn. Top-rated middle schools near 78704. Spanish immersion programs Cary North Carolina. AI assistants typically return three to seven named schools with one to two sentences of description each. The cited sources in this stage are dominated by GreatSchools, Niche, and the school's own websites.

Stage 3: Pairwise comparison. The parent asks the AI to compare two or three schools they are now considering. How does Trinity School compare to Spence. Acton Academy vs traditional Montessori. The pairwise comparison answer is the highest-leverage citation moment because the parent is converting from research to decision. AI models in this stage cite school websites, parent forum threads (College Confidential's K-12 equivalents, urbanbaby successors), and direct quotes from parent review aggregators.

Stage 4: Logistical detail. The parent moves to questions of fit and feasibility. What is tuition at the Wesley School. Does this school offer financial aid. What is the start time, the school calendar, the dress code, the lunch program. These are the queries where the school's own website becomes the dominant cited source — but only if the answers are exposed cleanly as text. Schools that bury this information inside PDF brochures or behind contact forms lose the citation to whichever source has the data as crawlable HTML.

Stage 5: Application calendar. Late-stage queries focus on deadlines, application requirements, and tour booking. Application deadline for [school name]. Open house dates spring 2026. ISEE testing schedule. AI models cite the school's admissions page directly here, which is the one stage where the school always wins the citation — but the conversion from citation to tour-booking depends on whether the call-to-action is structured for AI crawlers to surface.

The five-stage flow does not happen in a single session. Most parents conduct it over two to six weeks, with multiple separate AI sessions. The pattern has implications for which surfaces a K-12 operator needs to control at each stage.

Where AI Models Cite In K-12, By Query Type

The citation distribution varies sharply by the type of query the parent is asking. We segmented our 3,000-query audit by intent and tracked which sources appeared in the top three cited results for each query type.

A few patterns stand out. First, school websites are only cited reliably when the school has invested in operator-grade content infrastructure. Generic brochure-style sites lose the citation to aggregators every time. Second, state Department of Education sites are cited at much higher rates than most K-12 marketers realize — for any query touching standardized test data, the state DOE is in the top two citations roughly 60% of the time, which makes the DOE press release and data page another surface that operators should think about. Third, the camp category is the most fragmented, with regional parenting publications doing more citation work than national aggregators.

State Standardized Test Data As Citation Surface

One of the most under-discussed K-12 AEO surfaces is the state Department of Education website. AI models trust state DOE data as authoritative on standardized test results, accountability ratings, and demographic information. The implication for operators is that any school discussion that touches academic performance will pull from the state DOE page before it pulls from the school's own marketing.

The pattern is most visible in queries about public school quality. When a parent asks ChatGPT about the academic performance of a specific school, the response routinely cites the relevant state DOE accountability report alongside GreatSchools. The Texas Education Agency's TEA pages are cited in 78% of Texas public school queries we tracked. The California Department of Education's School Accountability Report Cards are cited in 71% of California public school queries. Florida's FLDOE School Grades pages are cited in 73% of Florida queries.

For private schools, the equivalent surface is the state's private school registry combined with any voluntary participation in standardized testing programs. Private schools that publish their average scores on the ERB, ISEE, or SSAT — alongside the percentile context — get cited materially more often in academic-performance queries than private schools that do not. The instinct to suppress test scores in private school marketing because they vary year to year is exactly inverted in the AI search era. Suppressed scores leave the citation slot empty, and AI models fill it with whatever third-party data is available, which is rarely flattering.

Charter networks have been the fastest to adapt to this dynamic. Success Academy publishes its New York State Assessment results prominently on the network's main site. KIPP publishes regional academic outcomes on each regional KIPP site. BASIS publishes AP exam pass rates with concrete numbers. These choices have produced disproportionate citation rates in AI search for academic-performance queries.

How Independent Schools Should Build Their AEO Surface

Independent schools — the roughly 30,000 private day and boarding schools in the U.S. — face the hardest version of the K-12 AEO problem. They compete with the GreatSchools-Niche citation moat, they cannot get test data from a centralized state source, and most of their marketing budgets do not support the editorial investment that wins citations. The schools that have figured out the playbook share a consistent set of choices.

1. Standardize the school profile across every aggregator. Claim and audit the school's pages on GreatSchools, Niche, Private School Review, BoardingSchoolReview, Findlay, and any local equivalent. Ensure the data points match: tuition by grade band, student-teacher ratio, accreditation bodies (NAIS, ISACS, NEASC), curriculum framework, religious affiliation if any, average class size, application deadlines. Inconsistent data across aggregators is one of the fastest ways to lose AI citation trust — models notice the inconsistency and discount all the sources. Schools that have invested in this standardization layer typically see their citation rate increase within 60-90 days of the audit.

2. Build a structured school profile page on the school's own site. This page should mirror the data exposed on the aggregators but with greater specificity. Include the matriculation list for the most recent graduating class (with college names spelled out, not just logos). Include average standardized test scores with percentile context. Include the exact accreditation timeline. Include the named curriculum framework with a substantive description of what it means in practice at this specific school. The page should be at a stable URL like /about/school-profile and render server-side. The schools doing this well — Sidwell Friends, Hewitt, Marlborough, Riverdale Country, the BASIS network — get cited as the source for their own data because the data is presented more comprehensively than the aggregators do.

3. Publish a parent-perspective content layer. Marketing copy is discounted by AI models. Parent voice is not. The schools winning AI citations have invested in named parent testimonials, alumni outcome stories, and tour-day recap pieces written in first person. These pages get cited as social proof in AI responses to fit-and-feel queries — what is the culture like at [school name], are parents happy at [school name], does [school name] have a strong arts program.

4. Maintain an admissions FAQ that mirrors actual parent queries. Build a comprehensive FAQ page that answers the specific questions parents ask AI assistants — tuition assistance criteria, sibling priority, faculty children admission, deferred admission, mid-year transfers, summer transition programs. This FAQ should follow the FAQ-format renaissance for AEO structure: question phrased as the parent would search, answer 80-150 words, self-contained.

5. Expose the school calendar, tuition, and admissions deadlines as structured HTML. Do not hide these in PDF brochures or behind contact-form gates. The schools that expose this data as crawlable text get cited in late-stage logistical queries. The schools that gate it lose the citation to whichever source — often a parent forum thread or a Niche review — has the data published.

6. Get coverage in regional education publications. Local journalism is one of the highest-leverage citation surfaces for K-12 because AI models trust independent third-party coverage more than they trust school marketing. Pitching coverage to outlets like Chalkbeat (which now has bureaus in seven states), regional NPR education reporters, and city-specific parenting publications builds the citation entity authority that compounds across queries.

The investment required to run this program well is meaningful — typically one full-time marketing role plus a content budget of $40,000 to $120,000 annually. Schools that have not historically staffed marketing at this level will find the shift uncomfortable. But the schools that have made the investment are winning the AI shortlist in their geographies, and the gap is widening every enrollment cycle.

The Tutoring Category: Varsity, Wyzant, Sylvan, Outschool

Tutoring AEO is a structurally different problem from school AEO. The queries are higher frequency, lower stakes per individual decision, and far more sensitive to local availability and pricing transparency. The competitive dynamics have produced a four-way concentration at the top of the citation rankings.

Varsity Tutors. Varsity dominates the head-term tutoring queries — best tutor in [city], math tutor [city], SAT prep tutor — with a 62% citation rate across our query audit. The dominance is built on three structural choices: comprehensive subject-and-grade landing pages with extraction-friendly content, transparent pricing displays (which AI models reward), and a large network of named tutors with individual bio pages. Varsity has effectively run the SaaS AEO playbook for documentation pages in the tutoring category, treating its programmatic landing pages as a primary citation surface rather than as SEO chaff.

Wyzant. Wyzant wins on a different surface: individual tutor profile pages. The site has roughly 80,000 active tutor profiles, each with reviews, hourly rates, and credentials. AI models cite Wyzant tutor profiles directly in answers about specific subject specialties or geographies. The citation rate of 44% is lower than Varsity's but the per-query depth is greater — Wyzant pages are quoted directly in answers, not just listed as a source.

Sylvan Learning. Sylvan's strength is local franchise coverage and Google Maps citation. AI Overviews that surface tutoring options for a specific city or zip code routinely cite Sylvan because the franchise model produces consistent local SEO signals across hundreds of metros. Sylvan's citation rate is lower nationally (28%) but its share-of-voice in geographies where it has a physical location is materially higher.

Outschool. Outschool has won the online enrichment category with a 54% citation rate for queries like best online classes for elementary kids, online chess classes for ages 8-12, and homeschool curriculum supplements. The win has been built on a category-specific surface — class listing pages with substantive descriptions written by named teachers — combined with strong third-party coverage in homeschool and gifted-education publications.

The implication for new tutoring brands trying to break into this category: the head-term citations are locked up, but the subject-specific and geography-specific long tail is still available. The brands gaining citation share in 2026 are vertical specialists — Mathnasium for math, Russian School of Mathematics for advanced math, IvyWise for high-stakes test prep, Outschool for enrichment. The path is to dominate the citations for a specific intent rather than to challenge Varsity head-on.

The Summer Camp Category Is The Most Fragmented

Summer camps are the most fragmented K-12 AEO category in our dataset. No single aggregator dominates the way GreatSchools dominates schools or Varsity dominates tutoring. The citation distribution skews heavily toward regional parenting publications, the American Camp Association (ACA) accreditation database, and individual camp websites that have invested in substantive content.

The American Camp Association reported in its 2026 industry update that 38% of first-time camp families used an AI assistant during selection, with peak usage in March and April when registration deadlines compress. The citation patterns for camp queries show three distinct dynamics:

Regional parenting publications dominate near-me queries. Mommy Poppins (New York), DC Refined (DC metro), Red Tricycle (multi-city), Bay Area Parent (Bay Area), and dozens of regional equivalents are cited at outsized rates for summer camp queries scoped to a specific geography. These publications publish annual camp guides that AI models index aggressively. Camps that appear in the regional publication's annual roundup gain a citation advantage that lasts the entire enrollment cycle.

The ACA accreditation database is the trust anchor. ACA-accredited camps are cited in 51% of sleepaway camp queries because the ACA accreditation signal is treated by AI models as an objective quality indicator. Camps that have not pursued ACA accreditation are visible to parents who already know them but invisible to parents discovering the category through AI search.

Camp websites win on substantive program descriptions. Camps that publish detailed daily schedules, named counselor bios, photo galleries with caption text (not just images), and parent testimonials by program type get cited materially more often than camps with marketing-copy websites. The format that works: a separate page for each program (junior boys, senior girls, CIT, specialty programs) with 600-1,000 words of substantive description, plus a parent-perspective FAQ covering the most common questions about homesickness, communication, food, and medical care.

The compression of the enrollment window is the underappreciated dynamic. Most camp families now book within 14 days of their AI search, which means the AI citation needs to convert quickly. Camps that have invested in fast-loading websites with clear registration calls-to-action capture the conversion. Camps with slow sites or buried registration pages lose families to better-converting competitors.

Charter Networks And Learning Pods

Two structural dynamics are reshaping how the non-traditional K-12 segment shows up in AI search.

Charter networks have built operator-grade AEO infrastructure. The large charter networks — KIPP, Success Academy, BASIS, Uncommon Schools, Achievement First, Great Hearts — have collectively become the most sophisticated K-12 AEO operators in 2026. Each network maintains a strong central site, regional sites for each metro, and individual school sites for each campus. The structure produces consistent entity signals across the network's footprint, and AI models cite charter networks in roughly 47% of relevant charter school queries. The most-cited network is KIPP at 38% national citation rate across charter queries, followed by Success Academy at 31% in New York and BASIS at 29% in their footprint markets. Charter networks have learned that AEO is an editorial discipline, and they staff it accordingly.

Microschools and learning pods are the long tail. The microschool and learning pod segment has grown to an estimated 750,000 students nationally according to a January 2026 report from the National Microschooling Center, but its AI visibility is hyper-local and fragmented. Microschool networks like Acton Academy, Wildflower Schools, Prenda, and KaiPod Learning have built network-level brand citation but the individual campus citations depend heavily on local parent Facebook groups, local news coverage, and word-of-mouth that gets transcribed into reviews. The networks that have invested in named-teacher content and parent-perspective testimonials on the network's main site (Acton Academy is the cleanest example) get cited in microschool queries. The independent microschools that have not are invisible to AI search.

For both segments, the marketing implication is the same: AI citation share is built through editorial infrastructure and third-party coverage, not through paid acquisition or local advertising.

What AI Models Get Wrong About K-12

The K-12 AEO category has higher rates of citation error than most other verticals because the data layer is fragmented and the school landscape changes quickly. The most common AI errors in our audit:

Outdated tuition figures. AI assistants routinely cite tuition figures that are one to three years stale. Schools that have raised tuition find that AI responses still quote the old number, which generates a credibility hit when families show up to the tour expecting the cited price.

Closed or renamed schools. AI models cite schools that have closed, merged, or renamed themselves. The training data lag is sometimes years, and the citation looks authoritative even when the underlying school no longer exists in that form.

Curriculum mischaracterization. AI assistants will describe a school as Montessori, IB, or classical when the school's actual program is different or has shifted. The school's own website typically has the correct framing, but if the aggregator data is wrong, the AI defaults to the aggregator.

Demographic data confusion. AI assistants frequently confuse demographic statistics across schools with similar names or in similar geographies. A query about one Spence school can return data about a different one. Schools with common names are particularly vulnerable.

Wrong leadership. Head of school turnover is high in private K-12, and AI assistants routinely cite former heads, former admissions directors, or former marketing leaders. The school's own About page is the canonical source, but if it is not updated promptly, the citation lag becomes a marketing problem.

For operators, the implication is that monitoring AI citations is a permanent operational function, not a one-time audit. The schools that have built recurring citation audits — quarterly is the minimum cadence, monthly is better — catch errors before they propagate and submit corrections to GreatSchools, Niche, and other aggregators on a regular basis.

The K-12 AEO 90-Day Playbook

For a K-12 operator — independent school, charter network, tutoring brand, summer camp, or microschool — looking to build AI citation share in the next 90 days, the prioritized sequence:

1. Run a baseline citation audit. Execute 50-100 queries across ChatGPT, Claude, Perplexity, and Gemini covering your category, geography, and the specific schools or programs you compete against. Document where you appear, where competitors appear, and which sources are being cited. This baseline is the foundation of everything else and typically takes a contractor 8-12 hours.

2. Claim and standardize every aggregator profile. GreatSchools, Niche, Private School Review, Findlay, ACA (for camps), Google Business Profile, Apple Maps, Yelp. Audit the data on each. Submit corrections. Ensure consistency across all surfaces. This is unglamorous work and it is the single highest-ROI move in K-12 AEO.

3. Build a structured school or program profile page on your own site. Mirror the data from the aggregators but with greater specificity. Render it server-side. Use a stable URL. Include matriculation lists, test scores in context, accreditation timelines, and substantive curriculum descriptions.

4. Publish a parent-perspective FAQ that mirrors real query language. Identify the 15-25 questions parents actually ask AI assistants about your category and your specific institution. Write 80-150 word answers, self-contained, in first-person or institutional voice. This single content investment typically drives more AI citations than any other on-page work.

5. Pitch coverage in regional and category-specific publications. Chalkbeat, Education Week, regional parenting publications, NPR affiliates, and any state-specific education trade press. The citations from third-party publications carry materially more weight in AI models than self-published marketing content.

6. Set up recurring citation monitoring. Use Profound, SerpRecon, or Bluefish to track your share of AI citations against competitors on a weekly or monthly basis. Many K-12 operators are using AI-specific monitoring for the first time in 2026 and finding that the data reframes the marketing measurement stack entirely.

7. Coordinate calendar, tuition, and admissions deadlines as crawlable HTML. Audit your admissions section for PDFs, gated content, and JavaScript-rendered components. Convert anything load-bearing for late-stage parent queries to clean HTML at stable URLs. Tour booking, application deadlines, and tuition by grade band should all be exposed as text.

8. Build a strategic third-party content pipeline. Identify the regional parenting publications, education podcasts, and category-specific trade press that AI models cite for your category. Build relationships with their editors. Submit pitches at the cadences they operate on. Track which third-party citations move your AI citation share.

The 90-day timeline is realistic for most operators with one full-time marketing person and a modest content budget. The compounding effects show up in the second and third enrollment cycles after the program is launched. Operators that have run this sequence consistently — typically the most sophisticated independent schools, the large charter networks, and the category-leader tutoring brands — are taking citation share from competitors every quarter the gap widens.

For an adjacent view on how the same dynamics play out in physical-presence local search across all categories, see the local AEO playbook for Google Maps and near-me queries. The K-12 category is one specific application of the broader local discovery shift, and the principles transfer across most service categories with geographic intent.

Takeaway: K-12 AEO in 2026 is a structured-data discipline first and a content marketing program second. The aggregators that dominate the citation layer — GreatSchools, Niche, ACA, regional parenting publications, state DOE pages — are not going away, and operators that treat them as part of their marketing surface (rather than as inert third-party noise) compound their citation share every quarter. The independent schools, tutoring brands, and summer camps that win the AI shortlist have built operator-grade infrastructure across school profile pages, parent-perspective FAQs, substantive program descriptions, and third-party coverage in regional publications. The enrollment cycles of 2026 and 2027 will widen the gap between operators who have built this infrastructure and operators who have not. The path is not glamorous and the work is not optional.