Schema.org for Program Pages: How Canadian Universities Get Cited by AI Systems

Why program-level markup is where Canadian universities win or lose AI citations

When a prospective student asks ChatGPT "best co-op engineering program in Ontario," the engine does not browse your homepage. It looks for structured, machine-readable signals at the program page level — signals that identify your program as a verifiable entity with specific attributes: credential type, duration, accreditation, tuition, and outcomes. Without those signals, your institution defaults to the pile of unstructured text that AI engines ignore.

Schools with structured Schema.org markup achieve an average of +12 percentage points in AI visibility compared to those without (Source: Skolbot GEO monitoring, 500 queries x 6 countries x 3 AI engines, Feb 2026). Yet across a sample of 120 Canadian institutions audited in early 2026, fewer than 14% had deployed EducationalOccupationalProgram markup on more than half their program pages. The gap is not technical complexity — the gap is awareness.

This guide covers the specific Schema.org types that matter for Canadian program pages, explains how OUAC application centre data intersects with markup strategy, and provides ready-to-implement JSON-LD examples calibrated to Canadian higher education conventions: credit hours (not ECTS), CAD tuition, OSAP eligibility, co-op work terms, and provincial quality assurance references.

For the foundational GEO framework, see our complete GEO guide for schools. For a broader checklist of what makes a program page AI-citation-ready, see How to get recommended by ChatGPT for "best school for [career]".

Choosing the right Schema.org type: EducationalOccupationalProgram vs Course

Canadian institutions frequently ask whether to use EducationalOccupationalProgram or Course. The distinction matters for AI citation precision.

EducationalOccupationalProgram is the correct type for full credential programs: bachelor's degrees, master's degrees, graduate diplomas, postgraduate certificates, and co-op program streams. It carries properties designed for credential-granting programs: educationalCredentialAwarded, programPrerequisites, applicationDeadline, occupationalCategory, and the offers block where tuition and funding eligibility live.

Course is appropriate for individual courses within a program — electives, required courses, and continuing education modules. Most Canadian universities should deploy EducationalOccupationalProgram at the program page level and, where resources permit, Course markup for individual course catalogue entries.

The practical rule: if it appears in your OUAC program listing and results in a provincially recognised credential, use EducationalOccupationalProgram. If it is a discrete learning unit within a program, use Course.

Canadian-specific properties you cannot skip

Several EducationalOccupationalProgram properties are particularly significant for the Canadian context and are frequently omitted:

• programType — Use "Bachelor", "Master", "Doctoral", or "Graduate Certificate". Canadian AI queries often filter by level of study.
• numberOfCredits + creditUnit — Canadian programs express credit load in credit hours or credits (not ECTS). Specify "creditUnit": "credit hours" explicitly. A four-year bachelor's is typically 120 credit hours.
• timeToComplete — ISO 8601 duration format. A standard four-year degree is "P4Y". An accelerated three-year bachelor's is "P3Y". A two-year master's is "P2Y".
• applicationDeadline — For OUAC-affiliated Ontario programs, the deadline must match your OUAC listing. For programs outside Ontario, reference the relevant provincial application centre deadline.
• offers.priceCurrency — Always "CAD". Specify whether the price is domestic or international tuition. Include OSAP eligibility in the description field of the Offer block.
• programPrerequisites — For Ontario programs: reference the OSSD (Ontario Secondary School Diploma) and OUAC application process. For programs in other provinces: reference the provincial equivalent and the relevant application centre (e.g., BCCAT for BC, Apply Alberta for Alberta).

Complete JSON-LD: a Canadian co-op program example

The co-op (cooperative education) program is one of Canada's most distinctive higher education offerings — and one of the most frequently queried by prospective students who type phrases like "best co-op business degree Canada" into AI engines. Co-op terms represent a concrete, verifiable, citable attribute. The Schema.org hasCourseInstance property is the right place to encode them.

Here is a complete EducationalOccupationalProgram block for a co-op bachelor's, ready to embed in a <script type="application/ld+json"> tag on the program page:

{
  "@context": "https://schema.org",
  "@type": "EducationalOccupationalProgram",
  "name": "Bachelor of Commerce (Co-op)",
  "description": "Four-year co-operative Bachelor of Commerce program with paid work terms integrated into the degree. AACSB-accredited business school, ranked in Maclean's top 10 for business.",
  "url": "https://www.university-example.ca/programs/bcomm-coop",
  "provider": {
    "@type": "EducationalOrganization",
    "@id": "https://www.university-example.ca/#organization",
    "name": "University Example",
    "sameAs": "https://www.university-example.ca"
  },
  "programType": "Bachelor",
  "educationalCredentialAwarded": "Bachelor of Commerce (BCom) — AACSB-accredited, provincially recognized degree",
  "numberOfCredits": "120",
  "creditUnit": "credit hours",
  "timeToComplete": "P4Y",
  "applicationDeadline": "2026-02-01",
  "hasCourseInstance": [
    {
      "@type": "CourseInstance",
      "courseMode": "onsite",
      "name": "Full-time stream with co-op work terms",
      "startDate": "2026-09-01",
      "endDate": "2030-04-30",
      "location": {
        "@type": "Place",
        "name": "Main Campus",
        "address": {
          "@type": "PostalAddress",
          "addressLocality": "Waterloo",
          "addressRegion": "ON",
          "addressCountry": "CA"
        }
      },
      "courseSchedule": {
        "@type": "Schedule",
        "description": "Six academic terms alternating with five paid co-op work terms (4–8 months total co-op experience)"
      }
    }
  ],
  "offers": {
    "@type": "Offer",
    "price": "8500",
    "priceCurrency": "CAD",
    "description": "Annual tuition (domestic students) — OSAP eligible, merit scholarships available"
  },
  "occupationalCategory": [
    "Financial Analyst",
    "Marketing Manager",
    "Management Consultant",
    "Accountant",
    "Business Development Manager"
  ],
  "courseMode": "onsite",
  "inLanguage": "en-CA",
  "programPrerequisites": {
    "@type": "EducationalOccupationalCredential",
    "credentialCategory": "Ontario Secondary School Diploma (OSSD) / provincial high school diploma — minimum 70% average, OUAC application"
  },
  "teaches": [
    "Financial accounting",
    "Marketing strategy",
    "Operations management",
    "Business analytics",
    "Organizational behaviour"
  ],
  "financialAidEligible": "OSAP (Ontario), Canada Student Loans Program (CSLP), institutional merit scholarships, bursaries"
}

What each block does for AI citation

• hasCourseInstance encodes the co-op structure as a typed object, not buried prose. ChatGPT and Perplexity can extract "five co-op work terms" as a discrete attribute and surface it in comparisons.
• provider with @id links the program to the institution entity defined on your homepage. Without this link, AI engines cannot attribute the program to your university — they treat it as an orphaned entity.
• occupationalCategory targets the "best university for [career]" query family. A student asking "best BCom program for becoming a marketing manager in Canada" needs to find "Marketing Manager" as a machine-readable field on your page, not buried in a paragraph about alumni success.
• financialAidEligible captures the high-intent query "which Canadian universities offer OSAP-eligible programs" — a query that filters enormously at the application stage.
• teaches captures skill-based queries: "business analytics degree Ontario", "organizational behaviour course co-op Canada."

Structuring OUAC-specific data in your markup

The Ontario Universities' Application Centre processes the majority of Ontario undergraduate applications. Its data structure differs fundamentally from UCAS (the UK equivalent): there is no UCAS points tariff, no personal statement scored by the application centre, and no clearing process. For AI engines trained on global higher education data, OUAC's mechanics are frequently misrepresented or omitted.

Your markup can correct this directly. In programPrerequisites, provide explicit, structured language:

{
  "@type": "EducationalOccupationalCredential",
  "credentialCategory": "Ontario Secondary School Diploma (OSSD)",
  "description": "Six Grade 12 U or M courses required; specific prerequisite courses vary by program. Application via OUAC (ouac.on.ca). Competitive admission — minimum average varies by year and applicant pool."
}

For programs outside Ontario, reference the correct provincial application centre:

• British Columbia: BC Council on Admissions and Transfer (BCCAT) / apply directly via institution
• Alberta: Apply Alberta (applyalberta.ca)
• Quebec: SRAM (Cégeps) / direct university application
• Maritime provinces: direct institutional application or Maritime Provinces Higher Education Commission (MPHEC) context

This level of specificity serves two functions: it gives AI engines the factual scaffolding to recommend your program accurately on province-specific queries, and it reduces hallucination risk where models otherwise fabricate UCAS-equivalent processes that do not apply in Canada.

The comparison table AI engines extract most reliably

AI engines extract structured tabular data at higher reliability than prose paragraphs. Include a markdown or HTML table on every program page. Here is a template calibrated to Canadian higher education queries:

Every row in this table is a potential AI citation anchor. A prospective student asking "does [university] co-op program accept international students" can receive a direct answer if your table includes the international tuition line. A student asking about accreditation receives a citation-ready fact with a year and renewal date.

Validating and maintaining your markup

Deploying JSON-LD once and ignoring it is the most common structured data mistake Canadian institutions make. Markup that references last year's tuition or an application deadline that has passed actively harms AI visibility: AI engines detect inconsistency between markup values and page content, and reduce trust in the entire domain's structured data.

Validation steps

1. Schema Markup Validator (validator.schema.org) — Check that your JSON-LD is syntactically correct and that all property names exist in the Schema.org hierarchy. Run this on every program page after deployment.

2. Rich Results Test — Google's tool verifies compatibility with enhanced search features. Errors flagged here often indicate properties that are present but incomplete.

3. Content consistency audit — Manually confirm that every value in your JSON-LD matches exactly what appears on the visible page: tuition, credit count, application deadline, accreditation status. Discrepancy between markup and page content is the primary cause of structured data penalty.

4. Quarterly refresh cadence — Set a calendar reminder at the start of each academic term to update: tuition for the new year, application deadlines for the next intake cycle, outcomes data when your institutional survey publishes, and accreditation dates when renewals occur.

Monitoring AI visibility changes

After deployment, run a set of ten program-specific queries through ChatGPT, Perplexity, and Gemini at baseline, at 30 days, and at 60 days. Queries should mirror the way prospective students phrase decisions: "best co-op business degree Ontario," "[university name] BCom OUAC deadline," "AACSB accredited business schools Canada." For a full KPI framework for tracking AI visibility, see our guide on ChatGPT and Perplexity visibility KPIs for schools.

Five markup mistakes Canadian institutions make

Mistake 1: Using Course where EducationalOccupationalProgram belongs

Course does not carry applicationDeadline, programPrerequisites, or educationalCredentialAwarded at the same semantic weight. A program page for a Bachelor of Science that uses Course markup instead of EducationalOccupationalProgram is under-typed — AI engines cannot extract the credential, the admission requirements, or the OUAC application link from a Course block with the same precision.

Mistake 2: Omitting priceCurrency or using USD

Every offers block on a Canadian program page must specify "priceCurrency": "CAD". AI engines operating globally will default to USD if currency is ambiguous. A prospect asking "how much does [program] cost at [university]" who receives a USD figure — even approximately correct — has been given inaccurate information that erodes trust in the AI engine's citation and in your institution.

Mistake 3: No hasCourseInstance for co-op programs

Co-op programs are one of Canada's strongest higher education differentiators internationally. Yet most co-op program pages describe the structure only in prose. Encoding co-op work terms in hasCourseInstance — with location, duration, and scheduling information — makes the co-op structure machine-readable and citeable. Without it, AI engines cannot distinguish your co-op program from a standard program with an optional internship.

Mistake 4: Accreditation listed as a string with no organisation reference

"accreditation": "AACSB" is useful but weak. A stronger signal:

"accreditation": {
  "@type": "EducationalOccupationalCredential",
  "credentialCategory": "professional accreditation",
  "name": "AACSB Business Accreditation",
  "recognizedBy": {
    "@type": "Organization",
    "name": "AACSB International",
    "url": "https://www.aacsb.edu/"
  }
}

The additional structure allows AI engines to cross-reference your accreditation claim with the AACSB public database — one of the verification behaviours documented in AI citation research. For a broader discussion of how accreditation signals drive AI recommendations, see our article on school accreditation and AI citations.

Mistake 5: Missing sameAs links on the provider

Your provider block should include sameAs links to your institution's entry in Universities Canada, your QS profile, your Maclean's profile, and your provincial ministry's degree-granting authority register. These cross-reference points are how AI engines verify that your institution is real, accredited, and independently recognised — not self-reported data from a single source.

How markup interacts with the broader LLM citation signals

Schema.org markup is necessary but not sufficient. AI engines that recommend programs combine structured data signals with three other signal types: content quality (verifiable figures in prose), external citations (mentions on trusted third-party sites), and entity consistency (the same facts appearing consistently across your website, your Universities Canada profile, and your OUAC listing).

A program page with perfect JSON-LD markup but no external citations — no mention in a Maclean's article, no Faculty of Engineering page on the accrediting body's site, no recent media coverage referencing the program by name — will still underperform relative to a competitor whose markup is less complete but whose program appears on twenty independent sources. For a full explanation of how these signals interact, see our guide on LLM signals used in school recommendations.

The practical implication: deploy your markup and then spend equal effort ensuring that every third-party profile — OUAC, Universities Canada, accrediting bodies — references your programs with the same attributes encoded in your JSON-LD. Consistency across sources is the multiplier.

FAQ

Does Schema.org markup on program pages affect OUAC application volumes?

Not directly — OUAC processes applications regardless of markup. The indirect effect is real, though: prospective students who first encounter your program in a ChatGPT recommendation and then visit your page are higher-intent than organic search visitors. They arrive with a specific question already answered and a higher likelihood of clicking through to your OUAC application link. The OUAC applicationDeadline field in your markup also captures the significant segment of students who ask AI engines "what is the OUAC deadline for [program]" — a high-intent query that precedes application.

Should we mark up French-language programs at bilingual institutions?

Yes. Use "inLanguage": "fr-CA" for French-language programs at bilingual institutions (University of Ottawa, Université de Moncton, University of Saint-Boniface). Add a separate EducationalOccupationalProgram block for each language stream rather than combining them into one block. AI engines handle language-specific queries separately, and a student asking "meilleur programme de commerce co-op Ontario en français" requires French-language markup to surface your program accurately.

How do we handle programs under provincial quality assurance review that are not yet fully accredited?

Include programPrerequisites and all available program attributes, but do not include accreditation fields for credentials that are pending review. Instead, use the description field in educationalCredentialAwarded to state the current status accurately: "Degree proposed for recognition under [provincial universities act] — quality assurance process in progress." Accurate representation of pending status is preferable to omitting the field entirely, and is substantially preferable to claiming accreditation not yet granted.

How long after markup deployment can we expect changes in AI citation rates?

Google rich results typically appear within two to four weeks of successful indexation. Changes in ChatGPT and Perplexity citation frequency take longer — typically four to eight weeks — because these models re-index content through RAG (retrieval-augmented generation) pipelines on their own schedules. Gemini, with tighter integration to Google Search, can reflect changes in two to four weeks. Brand-query improvements (a student searching specifically for your institution's program) tend to appear before generic-query improvements (appearing in "best co-op program in Ontario" results without a brand mention).

Schema.org EducationalOccupationalProgram markup is the most direct technical lever Canadian institutions have to convert their program pages from invisible text into citable entities. The co-op structure, OUAC integration, OSAP eligibility, and provincial accreditation data that make Canadian programs distinctive are precisely the attributes that AI engines are equipped to extract and surface — but only when encoded in structured data.