Get Recommended by ChatGPT: Programme Page BOFU Checklist

Why programme pages determine your ChatGPT visibility

When a prospective student types "best school for data science in the UK" or "top programme for UX design" into ChatGPT, the model does not run a search engine query. It synthesises an answer from patterns it has learned across thousands of web pages, drawing on sources it considers authoritative, structured and verifiable. The institutions that appear in those answers overwhelmingly share one characteristic: well-built programme pages.

29% of AI responses in the UK mention at least one school (Source: Skolbot GEO Monitoring, 500 queries × 6 countries × 3 AI engines, Feb. 2026). That means 71% of responses name nobody. For a Marketing Director or Director of Admissions at a private institution, the question is not whether ChatGPT recommends schools — it clearly does. The question is whether it recommends yours.

Programme pages occupy a decisive position in this dynamic because BOFU (bottom-of-funnel) queries are where recommendation happens. At this stage, an applicant already knows what career they want. They are asking a natural-language assistant to tell them which institution to shortlist. If your BSc Computer Science page cannot be parsed, cited or structured in a way ChatGPT can use, you will not appear — regardless of your rankings in the Guardian University Guide or the Complete University Guide.

This article gives you a practical 12-point BOFU checklist. It draws on the GEO (Generative Engine Optimisation) framework for UK schools and complements our broader analysis of what criteria AI engines use to recommend schools.

What ChatGPT analyses on a UK programme page

ChatGPT does not skim your page for brand mentions. It extracts structured signals. Understanding what it looks for lets you prioritise the changes that matter.

The model processes your page content against a set of implicit signals: outcome specificity (does the page state where graduates work?), structural clarity (are facts legible as discrete claims?), source authority (is the institution named alongside verifiable credentials like TEF ratings or QAA enhancement reviews?), and technical structure (is Schema.org markup present so the page declares its own subject matter explicitly?).

Table: Programme page elements and AI citation probability

Schools with structured Schema.org markup gain an average of +12 points in AI visibility (Source: Skolbot GEO Monitoring, 500 queries × 6 countries × 3 AI engines, Feb. 2026). That gap between a marked-up page and an unmarked one is the single largest technical lever available to admissions marketing teams.

The BOFU checklist: 12 optimisations for UK programme pages

Work through these in order. The first four are structural and have the largest impact; the remainder compound on top.

1. Add Course Schema.org markup with educationalCredentialAwarded

Every programme page needs a Course or EducationalOccupationalProgram structured data block. At minimum, include: name, description, provider (your institution as an EducationalOrganization), educationalCredentialAwarded (e.g. "BSc (Hons) Computer Science"), timeToComplete, and url. This is the single highest-impact change you can make. Without it, the model has to infer what your page is about. With it, you declare the answer explicitly.

2. State graduate outcomes with named employers

Replace "excellent career prospects" with specific claims: "87% of 2024 graduates secured employment within six months, with destinations including KPMG, the BBC and the NHS" (source your Graduate Outcomes survey, aligned with HESA reporting). Named employers and stated percentages are citable facts. Generic outcome language is not.

3. Publish your TEF outcome and link to the OfS register

The Teaching Excellence Framework award and your OfS registration are signals of institutional legitimacy that AI models have encountered repeatedly across UK higher education coverage. A one-sentence statement — "This programme is delivered by [Institution], registered with the Office for Students and holder of a TEF Silver award" — costs nothing to add and substantially strengthens your authority signal.

4. Include a marked-up FAQ section

Write four to six questions that mirror BOFU natural-language queries: "What jobs can I get with a [Programme] degree?", "What grades do I need to apply for [Programme] at [Institution]?", "Is [Programme] accredited?", "How long does [Programme] take?" Mark these up with FAQPage Schema.org. This is the mechanism by which your page content directly matches the queries that trigger AI recommendations.

5. Specify entry requirements as discrete facts

"Typical offer: ABB at A-level, including [subject]. UCAS tariff: 128–136 points. We also accept BTECs and Access to Higher Education Diplomas." Write it in plain prose with specific numbers, not a vague range buried in a table only a PDF renderer can parse. Applicants query entry requirements by programme constantly, and a page that answers this cleanly is far more likely to be cited.

6. Name the accrediting body and link to it

If your programme is accredited by a professional body — the Chartered Institute of Marketing, the BCS, the RIBA, the Law Society — state it explicitly on the page and link to the accrediting body's own website. External links to authoritative domains are a positive signal. They also give the model a verifiable claim to cite.

7. Add a salary benchmark from a credible source

Starting salary ranges from the Graduate Outcomes survey, the Prospects careers database or a named sector body give the page a data point that answers one of the most common BOFU queries: "How much will I earn with a [Programme] degree?" Even a single verified figure — "Median graduate starting salary: £28,500 (Graduate Outcomes 2023/24)" — raises the page's citation probability.

8. Publish a complete module list with brief descriptions

A semester-by-semester module list with 50–100 word descriptions does two things. First, it creates semantic depth: the page now contains hundreds of additional relevant terms that match specialised queries. Second, it answers a practical question applicants consistently ask. UCAS data confirms that module-level detail is among the top five pieces of information applicants seek before making a shortlist decision.

9. Embed authentic student or graduate testimonials with specific outcomes

"This programme helped me land a graduate scheme at Deloitte" is citable in a way that "life-changing experience" is not. Use full name and graduation year. Specificity signals authenticity, and authentic claims are what the model is looking for when constructing a recommendation.

10. Link to your ranking position with source and year

"Ranked 12th for [subject] in the UK — Complete University Guide 2026" is a verifiable claim. Link directly to the relevant Complete University Guide or QS World University Rankings subject table. Rankings cited without a source year or URL are treated as unverifiable. Those with a linked source are not.

11. Remove boilerplate and passive-voice copy from the page introduction

The first 200 words of your programme page receive disproportionate weight. Copy that opens with "Our innovative and dynamic programme prepares students for a rapidly evolving world" contains no extractable claims. Replace it with a direct declarative: who this programme is for, what qualification it leads to, and what graduates do. Front-load facts, not aspirations.

12. Ensure the page loads under 3 seconds on mobile and passes Core Web Vitals

AI crawlers reference the same publicly accessible content as search engine bots. A page that is technically slow or inaccessible due to render-blocking scripts is processed less reliably. The JISC digital infrastructure benchmarks for UK higher education recommend sub-3-second Time to Interactive on 4G. This is not merely a UX consideration — it affects whether your structured data is reliably read.

Common mistakes that keep UK schools invisible in AI responses

Publishing outcomes data only in PDF prospectuses. PDFs are processed less reliably than HTML by most AI crawlers. Any data that lives only in a downloadable prospectus — graduate employment rates, salary figures, accreditation details — is functionally invisible to generative models. Move it onto the page itself.

Hiding entry requirements behind a CRM form or a dynamic filter. If entry requirements only appear after a user selects their qualification type in a JavaScript dropdown, the static crawlable version of the page contains no entry requirements at all. The model cannot cite what it cannot read.

Conflating the programme page with a marketing page. Programme pages that are built primarily to convert through emotional copy — with hero videos, countdown timers and no factual content — are optimised for human psychology, not machine parsing. BOFU applicants and AI engines both need facts. A page that serves one serves both.

Omitting institution-level authority signals. A programme page that never names the institution's OfS registration status, TEF outcome or sector accreditations forces the model to infer authority. Institutions that state these credentials explicitly are cited more frequently, because the model has more to work with. For more on which signals matter most, see our analysis of content that gets cited by ChatGPT and Perplexity.

Treating GEO as an SEO task. Search engine optimisation and generative engine optimisation overlap but are not identical. SEO prioritises keyword density, backlinks and click-through rate signals that have no direct equivalent in an LLM. GEO prioritises structured data, factual specificity and verifiable claims. The teams responsible for programme pages need to understand both. Our guide to measuring AI visibility KPIs for schools covers how to track which approach is working.

FAQ

How long does it take for ChatGPT to reflect changes made to a programme page?

There is no definitive answer because OpenAI does not publish its crawl or training update schedule. In practice, for live ChatGPT Browse responses (where the model fetches current pages), changes can be reflected within days. For responses based on the model's training data, the lag is longer. Prioritise structured data and factual clarity now — it accumulates value over time regardless of update cadence.

Does a high position in the Guardian University Guide guarantee ChatGPT will recommend us?

Not directly. AI models incorporate ranking data as one signal among many. A school ranked 30th with excellent Schema.org markup, specific outcome data and a well-structured programme page will frequently outperform a top-10 institution with poorly structured content in AI recommendation responses. Rankings help, but technical structure amplifies them. For a full breakdown, see our article on AI recommendation criteria for schools.

Should we optimise every programme page or focus on flagship programmes first?

Start with your highest-traffic programme pages — those that already attract applicant interest and appear in your UCAS application data. Apply the full 12-point checklist to those first. The Schema.org implementation work can then be templated so that subsequent pages are cheaper to optimise. A phased approach produces measurable results faster than attempting to overhaul 200 pages simultaneously.

Is there a risk of over-optimising for AI at the expense of the applicant experience on the page?

No — because the checklist is aligned with what applicants want. Entry requirements, graduate outcomes, accreditation, module content and salary benchmarks are the information prospective students report needing most before shortlisting an institution. Optimising for AI citation and optimising for applicant clarity are the same task. A programme page that answers specific questions in plain English serves both audiences. The Educause review of digital student experience in higher education confirms that factual completeness — not marketing language — is the primary driver of page satisfaction among prospective undergraduates and postgraduates alike.

Programme pages are your institution's most powerful BOFU asset. They are the pages that applicants read when they are deciding whether to apply — and they are the pages that ChatGPT reads when it is deciding whether to recommend you. The 12 optimisations above are not theoretical; they are the specific structural changes that correlate with higher AI citation rates across Skolbot's monitoring panel of UK institutions.

For a broader view of where programme page optimisation fits in your overall GEO strategy, start with the GEO pillar for UK schools.