Schema.org Programme Page Markup: Structured Data That Gets Your Courses Cited by AI

Why Schema.org Markup Determines Whether LLMs Recommend Your Programmes

When a prospective student asks ChatGPT "best UK university for data science" or Perplexity "which business schools have AACSB accreditation", the AI does not return a list of links. It synthesises an answer from structured information it has processed — including Schema.org data from your programme pages.

Schools with structured Schema.org markup achieve an average of +12 percentage points in AI visibility (Source: Skolbot GEO Monitoring, 500 queries × 6 countries × 3 AI engines, Feb. 2026). In the UK, 29% of AI responses already mention at least one higher education institution — making this one of the most actionable improvements UK institutions can make right now.

The mechanism is not mysterious. ChatGPT, Perplexity and Gemini all rely on web crawls to build or refresh their knowledge. When a crawler encounters a programme page, it must infer what the page is about from prose. When it encounters a <script type="application/ld+json"> block, it reads a machine-structured declaration of exactly what the page describes, who offers it, what credential it awards, and what it costs. That declaration is processed with far greater reliability than page copy — and it is the information these engines draw on when constructing recommendation responses.

For the broader context on AI visibility for schools, see our guide to GEO for schools.

EducationalOccupationalProgram vs Course: Choosing the Right Type

The Schema.org vocabulary distinguishes between two types for educational offerings. Choosing the wrong one does not break validation, but it does send the wrong signal to AI engines that weight type declarations heavily when classifying pages.

EducationalOccupationalProgram: For Degree Programmes

Use EducationalOccupationalProgram for any offering that leads to a formal credential recognised by an awarding body or professional institution. In the UK context this covers:

• Undergraduate degrees: BSc, BA, BEng, LLB
• Postgraduate taught programmes: MSc, MA, MBA, LLM, MRes
• Integrated master's degrees: MEng, MPhys, MChem
• Professional qualifications: PGCE, Graduate Diploma
• Doctoral programmes: PhD, DBA, EdD

The EducationalOccupationalProgram type includes properties — programType, numberOfCredits, programPrerequisites — that are absent from the simpler Course type. These properties map directly to the signals LLMs use when answering queries such as "how many credits is a UK master's degree" or "what A-level grades do I need for [programme]".

Course: For Short Courses and CPD

Use Course for offerings that do not lead to a formal degree-level credential:

• Individual lecture modules
• Executive education short programmes (non-award-bearing)
• Continuing professional development (CPD) courses
• Summer schools and intensive programmes
• Online taster courses and MOOCs

If your institution offers both degree programmes and CPD, deploy EducationalOccupationalProgram on degree pages and Course on CPD pages. Do not use Course as a catch-all across your entire catalogue — an MSc page marked up as Course is type-mismatched against the queries that would otherwise surface it.

The 8 Schema.org Properties That Drive LLM Citations

The full EducationalOccupationalProgram vocabulary contains dozens of properties. The eight below are the ones that directly correlate with AI citation frequency, based on Skolbot's analysis of 500 queries across ChatGPT, Perplexity and Gemini. Implement all eight before considering optional extensions.

Here is a complete JSON-LD block for a UK BSc Business Management programme, followed by a property-by-property breakdown:

{
  "@context": "https://schema.org",
  "@type": "EducationalOccupationalProgram",
  "name": "BSc Business Management",
  "description": "Three-year undergraduate programme covering strategy, finance, marketing and operations. QAA-reviewed, with optional placement year and AACSB-accredited faculty.",
  "url": "https://www.university-example.ac.uk/courses/bsc-business-management",
  "provider": {
    "@type": "EducationalOrganization",
    "name": "University Example",
    "sameAs": "https://www.university-example.ac.uk"
  },
  "programType": "Bachelor",
  "educationalCredentialAwarded": "Bachelor of Science (BSc) in Business Management — QAA-reviewed, AACSB-accredited",
  "numberOfCredits": "360",
  "creditUnit": "credits",
  "timeToComplete": "P3Y",
  "applicationDeadline": "2026-01-15",
  "offers": {
    "@type": "Offer",
    "price": "9250",
    "priceCurrency": "GBP",
    "description": "Annual tuition fee — Student Finance England eligible"
  },
  "occupationalCategory": [
    "Business Analyst",
    "Management Consultant",
    "Marketing Manager",
    "Operations Manager",
    "Strategy Consultant"
  ],
  "courseMode": "onsite",
  "inLanguage": "en-GB",
  "programPrerequisites": {
    "@type": "EducationalOccupationalCredential",
    "credentialCategory": "A-levels: 120 UCAS tariff points (ABB)"
  }
}

Property 1: name — Use the full official programme title as it appears on your UCAS listing and your degree certificate. "BSc Business Management" not "Our flagship Business degree". LLMs match this field against the programme names students type, so consistency with UCAS and Discover Uni is essential.

Property 2: description — Write 50–150 words that pack in verifiable, specific facts: duration, key subject areas, accreditation status, notable features (placement year, study abroad, professional body recognition). Generic marketing language adds no signal value. Each specific fact — "QAA-reviewed", "AACSB-accredited faculty" — is a claim the LLM can cross-reference and cite.

Property 3: educationalCredentialAwarded — This is the highest-weight property for AI recommendation on qualification-specific queries. Include the full credential title plus any relevant quality marks in the string: "Bachelor of Science (BSc) in Business Management — QAA-reviewed, AACSB-accredited". When a student asks Perplexity "AACSB-accredited bachelor's in business UK", this field is where your page answers directly.

Property 4: occupationalCategory — List specific job titles graduates obtain, not generic aspirations. "Business Analyst", "Management Consultant" and "Operations Manager" are citable facts drawn from your Graduate Outcomes data. "Excellent career prospects" is not. This field drives recommendations on queries such as "which UK universities produce marketing managers" or "best degree for becoming a strategy consultant".

Property 5: offers with price and priceCurrency — Tuition fee information in structured data is the single most underused property by UK institutions. Of institutions analysed by Skolbot, fewer than 12% include fee data in their JSON-LD. Yet prospective students consistently search for fee information, and LLMs that have the data in markup can answer "affordable business degrees in London" with specific figures — citing your page.

Property 6: courseMode — Accepted values are onsite, online, blended. This field resolves a high-frequency query class: "online MBA UK" or "part-time master's London". An institution that marks up online on genuinely online programmes will be cited on queries that others miss entirely.

Property 7: programPrerequisites — Use EducationalOccupationalCredential with credentialCategory to state entry requirements precisely: "A-levels: 120 UCAS tariff points (ABB)". UCAS tariff point queries are among the most common BOFU searches. A programme page that answers this in structured data removes a friction point at the exact moment the applicant is deciding whether to apply.

Property 8: numberOfCredits — UK undergraduate programmes award 360 credits (120 per year); taught postgraduate programmes typically award 180 credits. Including this alongside creditUnit: "credits" lets LLMs answer credit-transfer and programme equivalence queries accurately — a growing query class as international students compare UK qualifications against their home country frameworks.

Marking Up UK Accreditations and Quality Ratings

Accreditations are the highest-trust signals in UK higher education AI responses. When Perplexity or Gemini answers "which UK business schools are AACSB accredited", it looks for the hasCredential property on your EducationalOccupationalProgram or EducationalOrganization blocks. Without it, your institution may not be cited even if you hold the accreditation.

Use the hasCredential property with an EducationalOccupationalCredential object that names the accrediting body and links to it:

"hasCredential": [
  {
    "@type": "EducationalOccupationalCredential",
    "credentialCategory": "accreditation",
    "name": "QAA Enhancement Review — Confidence",
    "recognizedBy": {
      "@type": "Organization",
      "name": "Quality Assurance Agency for Higher Education",
      "url": "https://www.qaa.ac.uk"
    }
  },
  {
    "@type": "EducationalOccupationalCredential",
    "credentialCategory": "accreditation",
    "name": "TEF Gold",
    "recognizedBy": {
      "@type": "Organization",
      "name": "Office for Students",
      "url": "https://www.officeforstudents.org.uk"
    }
  },
  {
    "@type": "EducationalOccupationalCredential",
    "credentialCategory": "accreditation",
    "name": "AACSB Accredited",
    "recognizedBy": {
      "@type": "Organization",
      "name": "AACSB International",
      "url": "https://www.aacsb.edu"
    }
  }
]

The accreditations that carry the most weight in UK LLM responses, by query frequency:

• QAA — The Quality Assurance Agency Enhancement Review outcome. State "Confidence" or "Limited Confidence" precisely. Do not mark up an expired review.
• TEF — Teaching Excellence Framework Gold, Silver or Bronze. Include the award year if it differs from the current TEF cycle. Link to the OfS register entry.
• Professional body recognition — CIM (marketing), CIPD (HR), CIMA/ACCA (accounting), BCS (computing), ABET (engineering), AMA (architecture). Each body publishes an accredited programme list that LLMs cross-reference.
• Business school triple accreditations — AACSB, EQUIS, AMBA. Any combination of these is a strong recommending signal on MBA and business master's queries.
• Professional licensing bodies — GMC (medicine), the Law Society (law), Engineering Council (engineering). These are mandatory for LLMs to route "how do I qualify as a [profession] in the UK" queries correctly.

One critical rule: mark up only current, active accreditations. An LLM that detects a contradiction between your markup and the accrediting body's own published list — for instance, a lapsed EQUIS accreditation still declared in your JSON-LD — will treat the inconsistency as a trust signal failure across your entire domain.

Graduate Outcomes: The Most Cited Data Point

When prospective students ask "what can I do with a [degree] from [university]", LLMs prioritise institutions that provide specific, verifiable graduate data. Vague outcome language — "graduates go on to successful careers" — cannot be cited. Specific data can.

The primary source for UK graduate outcome data is the Graduate Outcomes survey, run by HESA. Use its figures in your occupationalCategory and in your page prose, cited with the survey year:

"occupationalCategory": [
  "Data Scientist",
  "Business Intelligence Analyst",
  "Machine Learning Engineer",
  "Data Engineer",
  "Product Manager (Technology)"
],
"description": "87% of 2023/24 graduates in employment within 15 months (Graduate Outcomes Survey 2024/25). Primary destinations: technology, financial services, consulting."

Three practices that raise citation frequency on outcome queries:

1. Name employers, not sectors. "Graduates have joined KPMG, Rolls-Royce, the NHS and Sky" is citable in a way that "graduates work across finance and technology" is not.
2. State the salary benchmark. Median starting salary from the Graduate Outcomes survey, with the survey year, is the highest-value single data point you can add to a programme description.
3. Align job titles with Standard Occupational Classification codes. The Graduate Outcomes survey uses SOC codes. Using the same job titles in your occupationalCategory creates a verifiable link between your markup and the published dataset that LLMs can cross-reference.

Testing and Validating Your Structured Data

Before publishing any JSON-LD markup, run it through validation in sequence. Syntactically correct markup that is semantically inconsistent — for example, a timeToComplete value of P4Y on a three-year programme — is actively harmful: it gives LLMs contradictory signals and erodes trust across the page.

The workflow: Schema.org Validator first (vocabulary check), Rich Results Test second (Google eligibility), then publish and monitor Search Console for errors that only appear in production — typically caused by CMS output escaping characters inside the JSON-LD block.

Schedule a quarterly audit in Search Console under "Enhancements". JSON-LD markup frequently breaks silently after CMS updates, template changes or fee updates that alter the page structure.

Common Mistakes UK Institutions Make

1. Using generic language in occupationalCategory instead of specific job titles. "Excellent career prospects" or "careers in business" are not valid occupationalCategory values and produce no AI matching signal. Use the exact job titles from your Graduate Outcomes data: "Financial Analyst", "Supply Chain Manager", "Digital Marketing Executive".

2. Omitting UCAS tariff points from programPrerequisites. Entry requirements are among the most queried BOFU signals. An institution whose markup includes "A-levels: BBB, 120 UCAS tariff points" will be cited on entry requirement queries; one that omits this will not.

3. Not including tuition fees in offers. The prospect query "affordable MBA London" or "cheapest law degree UK" is answered by LLMs using price data in structured markup. An institution that omits fees forces the LLM to ignore it on fee-sensitive queries — a significant portion of BOFU search volume.

4. Using Course type for degree programmes. Course lacks programType, numberOfCredits and programPrerequisites. An MSc marked up as Course is missing the properties that distinguish it from a CPD module in the LLM's classification.

5. Inconsistent naming between Schema.org name and UCAS or Discover Uni listing. If your UCAS entry reads "Business Management with Marketing BSc (Hons)" but your JSON-LD name reads "BSc Business Management", the LLM may fail to match the two as the same programme. Use the canonical UCAS title.

6. Missing hasCredential markup for professional body recognition. An institution with CIPD-recognised HR programmes that does not declare this in structured data will not appear when Gemini answers "CIPD-recognised HR master's UK". The accreditation exists; without markup, it is invisible to the model.

FAQ

Does every programme need its own Schema.org markup?

Yes — one EducationalOccupationalProgram block per programme page. LLMs process pages individually; a single generic block on a "courses overview" page does not trigger programme-specific citations. When a student asks "best data science MSc in the UK", the LLM needs to encounter the markup on the MSc Data Science programme page specifically.

Can our CMS generate the JSON-LD automatically?

Most major CMS platforms support JSON-LD templates. WordPress with RankMath Pro or Schema Pro, Drupal with the Metatag or Schema.org Blueprints module, Craft CMS with custom field mappings, and headless CMS platforms via custom API integrations can all generate programme-level JSON-LD from content fields. Validate the output with the Rich Results Test after implementation — auto-generated markup frequently misses optional but high-value properties such as offers and occupationalCategory.

How long before AI systems pick up our new markup?

Google typically crawls and processes structured data within one to four weeks of publication. For Google AI Overviews, the effect on citation frequency can be observed within the same window via Search Console. For ChatGPT and Perplexity, the lag depends on their crawl and update cycles — typically one to three months for base model knowledge, but faster for retrieval-augmented responses that fetch live pages. The markup improves crawlability from day one regardless of the downstream update cycle.

Does Schema.org replace other GEO signals?

Structured data amplifies your existing authority signals rather than replacing them. An institution with strong Wikipedia presence, linked authoritative external mentions, and consistent NAP (Name, Address, Phone) data across the web will see a larger uplift from Schema.org than one without these foundations. For the full picture of what LLMs evaluate, see our article on the 15 signals LLMs use for school recommendations.

How do we know if our programme pages are being read by AI systems?

The most direct method: query ChatGPT, Perplexity and Gemini directly about your programmes and your institution. Use queries that mirror what prospective students ask — "best [subject] degree [city] UK", "[your institution] [programme] fees", "is [institution] AACSB accredited". For systematic monitoring across query sets and engines, see our AI visibility audit tools comparison for higher education.