Every week another agency deck arrives with "AEO" on slide four. Somewhere between "SEO" and "paid social" now sits a new box called "answer engines" — sometimes broken into ChatGPT, Perplexity, Google AI Overviews and a couple of others — with a proposed monthly retainer bolted to it. The pitch is that you now need someone to run this the way you have someone running your Google Ads account.
This is the wrong shape for the problem. AEO is not a channel. It is a distribution layer sitting on top of every other thing an institution publishes, and treating it as a separate workstream produces the exact outcome those agency decks are supposedly solving for: your content optimises for one surface and vanishes from the others.
What a channel actually is
A channel is a distribution surface with a distinct account, a distinct budget, a distinct set of creative assets and a distinct set of measurable actions on the far side of it. Google Ads is a channel because there is an account, there is a spend line, there is a set of ads, and there is a click-and-conversion event. Meta is a channel for the same reasons. LinkedIn organic is a channel — no spend, but everything else holds.
An answer engine is not that. There is no ChatGPT Ads account for education institutions. There is no Perplexity ad server accepting your creative. There is no bid on "best Christian university in Brisbane" that puts your institution into the answer. The distribution is happening — but it is happening as a consequence of what the assistant's underlying model has crawled, indexed, weighted and decided to cite, based on the same signals that already drive organic search plus a few new ones (structured data, citation-worthy formatting, entity clarity).
Which is why the "AEO retainer" pitch is a category error. You cannot have an agency operating your answer-engine account. There is no account.
What is actually going on
Three things:
- The same content is being consumed by more surfaces than it was two years ago. A well-structured course page that answers a specific question is being drawn on by Google search, Google AI Overviews, ChatGPT, Perplexity, and — increasingly — the AI features being embedded into Bing, Meta and TikTok. One asset, five to seven surfaces.
- The signals that make content citable have expanded. Schema.org structured data, direct Q&A formatting, clear entity naming, verifiable statistics with sources — all of these were "SEO best practice" that most institutions half-implemented for the last decade. They now carry meaningfully more weight because they are what LLMs preferentially extract from.
- The measurement is genuinely harder. A prospective student who reads about your institution inside a ChatGPT answer and then visits your site directly the next day shows up in your analytics as a "direct" or "organic branded" session, not as an "AEO" one. There is no referrer. This is not a bug that gets fixed — it is a structural feature of how the distribution works.
What to do instead
Stop treating AEO as its own workstream. Fold it into the existing content and technical SEO practice, on the following basis:
Every course page is an answer to a specific question
Not a category listing. Not a marketing landing page with the course details buried in the middle. An actual, direct answer to the question a prospective student is asking their AI assistant this week — with the answer visible in the first two hundred words, structured with proper headings, and marked up with course-related schema (Course, EducationalOccupationalCredential, EducationalOrganization).
The FAQ section is the AEO engine
Not because FAQs are magic, but because FAQPage schema plus a well-written question-and-answer pair is exactly what LLMs preferentially cite when generating a response. If your institution's FAQ set is written to genuinely useful, sector-specific questions rather than "How do I apply?" filler, the AEO work is largely done.
Named entities matter
The institution's name, the campus locations, the course names, the accreditation body, the sector-specific terms (CRICOS provider number, TEQSA registration, ATAR minimums) — these are the entities LLMs use to build a citation-worthy answer. Bury them in prose and the assistant guesses. Surface them cleanly and the assistant quotes.
Statistics with sources beat statistics without
An unattributed "our graduates report high employment outcomes" is worth close to zero. A specific "87 per cent of our nursing graduates were employed in a clinical role within twelve months of graduation, per the 2025 Graduate Outcomes Survey" is worth an enormous amount — LLMs will pull the specific number, cite it back, and increasingly link to the source.
The measurement question
Because AEO has no attribution model, the measurement is not "AEO-attributed conversions." That number does not exist and will not exist. The measurement is:
- Direct and organic-branded traffic volume, quarter-over-quarter
- Manual sampling of AI assistant answers to your top-of-funnel prospect questions ("best Christian university in Brisbane", "Australian tertiary providers with online theology") — with a documented record of whether your institution is mentioned, favourably, with correct information
- Enrolment enquiry origin as declared by the applicant ("how did you first hear about us") — increasingly a useful signal, because prospective students are now naming ChatGPT and Perplexity explicitly
None of these are as clean as "clicks and conversions from source X." That is the honest answer. The compensating factor is that the volume from these surfaces is compounding — an institution that gets this right in 2026 will be the answer to the sector-relevant question for the next several years, because the underlying models keep re-crawling and keep weighting the same signals.
What this changes about agency selection
If you are being pitched "AEO as a service" at a monthly retainer separate from your content and technical SEO, that is worth asking about. It may be a well-intentioned attempt to give this new discipline a home in the pricing structure. It may also be a category error being sold as a product.
The better structure — and the one used at EduGrow Vantage — is a single Search & Answer practice that treats SEO, AEO and GEO as three surfaces of the same work. Same audit. Same content restructure. Same schema deployment. Different measurement mix per surface, but one workstream doing the work.
The prospective student who searches Google, is answered by ChatGPT, and is recommended by Perplexity — sometimes in the same evening — is not experiencing three channels. Neither should your practice be organised around three.