What is AEO? A practical beginner’s guide to answer engine optimisation

AEO is a working term, not a separate certification

AEO usually means answer engine optimisation: making information easier for systems to retrieve, understand, and use when answering a question. You may also hear GEO, or generative engine optimisation. People use these terms differently. Neither label guarantees a particular platform will mention your business.

SEO focuses broadly on discoverability in search; AEO draws attention to answer-oriented experiences, including AI-generated responses. The practical work overlaps: useful accessible pages, clear organisation, credible evidence, and consistent information.

Begin with questions real customers ask

Collect ten important questions from conversations, support, search queries, and sales objections. Choose one you can answer well. Write a short direct answer, then explain the method, context, examples, and exceptions. Use headings a reader can understand without needing the rest of the article.

For example, “How much can I afford to pay per click?” deserves a formula, a worked fictional scenario, and a warning that conversion assumptions need evidence. A confident one-line answer without assumptions can mislead both humans and systems.

Make important facts consistent and accessible

• Keep business name, offer scope, contact information, product details, and prices consistent wherever you maintain them.
• Use descriptive headings and readable HTML text rather than putting the entire answer in an image.
• Identify authorship, sources, review dates, and the basis for factual claims.
• Provide original examples or evidence where you can, and make the limitations visible.
• Use structured data only when it accurately matches the content and the platform’s documented requirements.

Google says there are no extra technical requirements or special optimisations necessary for its AI search features beyond the relevant search foundations. AEO should not become an excuse to ignore basic crawlability or the quality of the answer.

Understand mentions, citations, and visits

A mention names your brand. A citation points to a source. A referral visit occurs when someone follows a link and your measurement can observe the arrival. These are different events. A cited page may receive few clicks; a customer may encounter an answer and visit later through another route.

Track a small, stable set of customer questions in a dated visibility log if you want to sample answers. Record platform, wording, date, location or account context where relevant, the answer, and any cited URL. Results can change. This is a sample, not a complete market-share measurement.

Measure customer usefulness as well as visibility

Inspect identifiable AI referral sources in analytics when available, landing pages, useful actions, and qualified inquiries. Add an optional “How did you hear about us?” question to your business process where appropriate. Responses and referral labels are imperfect; use them as complementary signals.

Do not infer that no recorded referral means no influence. Equally, do not attribute every new inquiry to AI because you recently updated an article. Keep observations and interpretations separate.

Try a four-week improvement cycle

Week one: choose the questions and audit the current answers. Week two: improve two pages with clear explanations, evidence, examples, and accurate business facts. Week three: check technical access and share the resources with relevant people. Week four: review sampled answers, observed referrals, and customer feedback.

This is an operating exercise, not a promise of citations in four weeks. Avoid buying guaranteed mentions, fabricated reviews, or a supposedly mandatory AI file. Emerging conventions such as llms.txt are not a universal requirement for appearing in answers. Make the pages useful enough to justify the work independently.