AEO in 2026: how brands get cited by ChatGPT, Perplexity and Google AI
In short
Answer engine optimization means making your brand the source an AI assistant quotes. Write one clear answer per page, keep your entity facts consistent, mark up products and people with schema, and publish a machine-readable summary. Then track which prompts surface you and fix the gaps.
Search did not disappear, it moved. More buying questions now get answered inside ChatGPT, Perplexity and Google's AI panels before anyone opens a website. If those answers never mention you, the shortlist is written without you.
What AEO actually changes
Answer engine optimization is often described as SEO with a new name. The foundations are the same: crawlable pages, clear structure, credible sources and real expertise. What changes is the format of the answer.
Classic search returns a list of links. An answer engine reads sources, assembles a response and attributes it to the few brands it trusts on the question. Being in the index is not enough; you have to be quotable.
That shifts the work from ranking pages to shaping sentences. The brands that get cited state a clear position in language a machine can lift without rewriting it.
The brands that adapt fastest keep doing the SEO basics and treat the answer layer as a new output format for content they should already be publishing. No shortcut replaces the fundamentals.
Write answer-first, then expand
Every important page should open with the answer. Forty to sixty words that a person reads in one breath and an assistant can quote in one block. Then the detail, the nuance and the exceptions.
Write headings as the questions buyers actually type. 'How much does a website cost' beats 'Pricing philosophy'. 'Do 3D tours work for luxury homes' beats 'Immersive media'.
The same discipline helps human readers, which is the point. Content that is easy to quote is content that is easy to understand.
- Answer in the first 60 words, then expand with context, method and limits.
- One question per heading, phrased close to natural language.
- Short paragraphs, concrete nouns, no throat-clearing.
- State what you do, for whom, where and how to start.
Make the entity unambiguous
Assistants build answers from entities: people, brands, products and places they can identify with confidence. If your brand appears under three spellings, with two addresses and no clear description, you are harder to cite.
Audit how your brand reads across the web, then make it boringly consistent. This is unglamorous work, but it decides whether an assistant treats you as a source or as noise.
- One canonical description of your business, reused on your site, profiles and directories.
- Named people with real bios, roles and credentials, linked to the brand.
- Consistent names, addresses and categories everywhere they appear.
- SameAs links that connect your site to your verified profiles.
Schema, llms.txt and the machine-readable layer
Structured data gives machines a clean version of what your page already says: Organization, Article, Product, Person, LocalBusiness, FAQPage. Google's rich results come and go, but the markup still helps systems parse facts about you.
The llms.txt convention is younger. A short file at the root of your site points assistants to the pages that matter most. It is cheap to publish, easy to keep current and not yet a guaranteed shortcut. Treat it as a clarity exercise, not a hack.
This layer should mirror what is visible on the page. Markup that contradicts the text, or ratings you cannot substantiate, is worse than no markup at all.
What makes a brand quotable
Answer engines quote sentences they can defend: definitions, comparisons, named processes and numbers with a source. Vague marketing copy gives them nothing to hold onto.
Publish content only you can write. A studio that shows how it structures a property launch, or how it prices a build, gives assistants a reason to cite it over a generic listicle.
Proof matters too. Named people, real projects and clear contact details make you a safer citation. We describe our own work the same way: e-commerce for Gekko Beers, paid media for Passion d'Ames, brand and campaign exploration for Geothermal.
How to measure AEO
AEO has no single dashboard. Measure it like a panel: a fixed set of prompts, run on a fixed schedule, across the assistants your buyers actually use.
Track when you are mentioned, when you are cited, when you are quoted with a link, and who appears instead of you. Keep a simple log; the pattern matters more than any single answer.
Server logs help too. AI crawlers identify themselves, and you can see whether they are visiting the pages you want them to read.
- A prompt panel: 30 to 50 questions your buyers ask, run monthly, results logged.
- Crawler logs: which AI bots visit, which pages they read, which they ignore.
- Referrals: sessions and conversions from assistant surfaces, tagged and reviewed.
- Citation share: how often you appear versus the three competitors that matter.
Partly. The foundations overlap, and you should fix SEO before chasing citations. AEO changes the packaging and the measurement: you optimise sentences and entities, and you track mentions rather than positions.
It is a convention, not a standard, and no assistant has promised to read it. Publish one because it is cheap and it forces clarity, but keep investing in the pages assistants already crawl.
Ask it. Run a fixed set of buyer questions across ChatGPT, Perplexity and Google's AI surfaces every month, log who is mentioned, and treat changes as signals rather than proof.
Your homepage and service pages: a plain answer, a clean entity description, correct schema and a contact route. Most brands find gaps there before they need anything advanced.