AEO questions, answered
Straight answers to the questions businesses ask us about Answer Engine Optimisation. Each answer is complete on its own; where a question deserves a full page, we link it. Sourced from real conversations and from the unanswered-questions loop we run on our own site, the same system we run for clients.
The basics
What is Answer Engine Optimisation?
Answer Engine Optimisation (AEO) is the practice of getting a business cited in the answers AI engines give, ChatGPT, Google AI Overviews, Perplexity and Gemini, when buyers ask for a recommendation or a shortlist. Where SEO competes for a position in a list of links, AEO competes to be the source the answer is built from. The work is concrete: answer-first content, per-page schema, a consistent entity footprint, and coverage in the languages buyers actually use.
Is AEO the same as GEO?
Yes, for practical purposes. GEO (Generative Engine Optimisation) is another name for the same discipline, coined in academic research, and the industry has not settled on one term. The tactics are identical either way. We say AEO because it names the outcome a buyer cares about, being the answer, and we mirror whichever term a client's team already uses.
How is AEO different from SEO?
SEO optimises to rank a page in a list of results a person scrolls. AEO optimises to be cited inside the single answer an AI engine writes instead of that list. SEO wins traffic; AEO wins the recommendation, and most businesses now need both because buyers use both. The tactics overlap, clean structure and useful content help each, but the targets differ enough that you can rank well and still be missing from every AI shortlist.
Do people really ask AI for business recommendations?
Yes. Asking an AI engine for a shortlist, best clinic for a procedure, which dealer to trust, who to hire for a job, has become a normal first step in buying research, the way a search used to be. The behaviour is easy to confirm yourself: ask an engine the question your best customer would ask and it returns named businesses, not a list of links.
The shift matters most for high-consideration purchases, where buyers research heavily before contacting anyone. There, the shortlist is often formed inside the AI conversation before any website is visited, which is exactly why absence from the answer is invisible to the business losing the enquiry.
Which AI engines matter for AEO?
Four surfaces carry most of the buyer intent worth winning: ChatGPT, Google AI Overviews, Perplexity and Gemini. ChatGPT is the default first stop for a large share of buyers. AI Overviews sits above the classic Google results and reaches people who never chose an AI tool at all. Perplexity cites its sources inline by design. Gemini is grounded in the same signals as Overviews and is growing quickly in Southeast Asia.
Share shifts between them, and the concentration is real: one 2026 analysis of 6.77 million AI-referred sessions found 92% came from ChatGPT, with the caveat that AI activity inside Google's own results is measured separately and is likely larger still. That concentration is why we track citations across all four surfaces rather than optimising for one engine's current lead.
How citation works
How does an AI engine decide which businesses to cite?
It cites sources it can read, understand and trust. Reading means crawler access and real HTML text. Understanding means direct answers near the top of the page, and schema that states who you are and what you sell. Trust means a consistent entity footprint: the same description of the business on your site and across the profiles, directories and third-party sources the engine already knows.
No single input earns a citation on its own; the engines weigh them together. The practical consequence is that a gap in any one of them, blocked crawlers, no schema, brochure copy, an inconsistent identity, is enough to keep a business out of an answer it should be in.
Can I pay to be included in AI answers?
No. Citations are not ad placements, and there is no product that inserts a business into a synthesised answer. Engines select sources they can read, verify and trust, which is also what makes a citation durable once earned: it does not stop when a campaign budget does. The productive version of this question is diagnostic, why an engine cites a competitor and not you, and that has six concrete, checkable causes.
Why does ChatGPT recommend my competitor and not me?
Because in some concrete, checkable way your competitor is legible and verifiable to the engine and you are not. The causes are a short list: blocked AI crawlers, no structured data, content that never answers a question, a thin citation footprint, single-language coverage, or pages a machine cannot read. Every one has a two-minute self-check, so the full diagnosis is worth running before assuming anything mysterious is happening.
Does blocking AI crawlers in robots.txt matter?
Yes, more than any other single technical factor. A robots.txt that disallows GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot or Google-Extended makes the site structurally invisible to those engines: you cannot be cited from pages that were never crawled, whatever the content quality. Many businesses blocked AI crawlers by default in an earlier policy decision and forgot. Check yourdomain.com/robots.txt for those names next to a Disallow rule; the full crawler list and the fix are in the guide.
Does schema markup actually influence AI answers?
Yes, as one input among several. JSON-LD schema (Organization, Service, FAQPage, LocalBusiness) tells an engine what a page is, who the business is and how the facts relate, instead of leaving it to infer everything from prose. It makes you legible; it does not make you citable on its own. Schema wrapped around a page with no useful answer earns nothing, and useful answers without schema make the engine work harder than it needs to. The pairing is the point: direct answers in the copy, schema confirming the facts underneath.
What is llms.txt and do I need one?
llms.txt is a proposed convention: a short markdown overview of your site, hosted at /llms.txt, written for AI crawlers rather than people. It is an emerging practice, not a confirmed input to any engine, so treat it as cheap hygiene rather than a citation driver. Host a short one generated from your real pages, then move on to the work with more evidence behind it: answers, schema, entity clarity. We host one ourselves and generate one for clients as an audit deliverable.
Do AI engines read content in Thai, Chinese and Arabic?
Yes. The major engines read and answer fluently in all three, and buyers prompt in their own language. What an engine cannot do is cite content that does not exist: if a buyer asks in Chinese and your site only answers in Thai or English, the engine quotes a competitor who is quotable in the buyer's language. That is why multilingual answer coverage is an AEO input rather than a nice-to-have, especially for businesses serving international buyers across Southeast Asia.
Doing the work
What does an AEO audit actually cover?
Whether AI engines can find, understand and trust your business, measured page by page. In practice that means crawler access, schema coverage, whether pages answer real buyer questions, your citation position against named competitors, and language coverage, returned as a scored report with the fixes attached: per-page schema, content plans and a prioritised roadmap. The audit page covers the deliverables in full, and a free snapshot there scores your homepage instantly if you want a first read.
How long until AEO shows results?
It varies by engine and starting point, and anyone promising a fixed timeline is guessing. The honest shape: crawler and schema fixes are re-read by engines within days to weeks, while the citation footprint that builds trust accumulates over months. On-page changes are verifiable immediately; citation movement shows up in sampling later. Distrust confident dates in either direction.
Can AEO results be measured?
Yes, honestly framed. Citation tracking asks the engines the questions your buyers ask and records which businesses get named; we run multiple passes and average them, and we call it sampling because that is what it is, not real-time monitoring. On-page inputs (crawler access, schema validity, answer-first structure) are verifiable immediately and objectively. What measurement should never do is invent precision: a citation share is an estimate with a method behind it, and the method should be stated.
Should I stop doing SEO?
No. SEO and AEO are different jobs done on the same asset: strong pages, clean structure and genuinely useful content serve both. Traditional search still sends real traffic, and many AEO inputs reinforce rankings. AEO is an addition that covers the growing share of buyers who read one AI answer instead of scrolling a list, not a replacement for the visibility you already have.
Can I do AEO myself?
Parts of it, yes, and we publish the material to do it. The audit checklist shows the nine areas a serious review covers, and the citation guide walks through the fixes: answer-first content, schema, entity consistency, crawler access. A competent team can get real gains from those two pages alone.
The parts that are hard to do in-house are multilingual answer coverage, measurement with an honest method behind it, and sustained iteration as the findings keep changing. That is usually the line where doing it yourself stops making commercial sense.
What is an on-site AI agent and why does it matter for AEO?
An on-site AI agent answers visitor questions from your own content, grounded in a single knowledge base, in the visitor's language. For AEO it matters twice. First, it converts the visitors your citations bring: a buyer who arrives with a question gets an answer instead of a contact form. Second, every question the agent cannot answer is logged as a content gap, and that unanswered-questions loop feeds exactly the content AI engines want to cite: real buyer questions, answered directly. It turns a site from a static brochure into a system that keeps finding its own gaps.
Thailand and Southeast Asia
Why does AEO matter specifically for businesses in Thailand?
Because the buyers worth winning prompt AI in several languages, and most business sites here can answer in only one. Thai buyers ask in Thai. International buyers, medical-tourism patients, regional B2B, expats, ask in English, Chinese or Arabic.
The language gap is measurable: Google's Gemini Southeast Asia Report 2026 found nearly 70% of prompts in the region are submitted in native languages, with Thailand at 87%. An engine answering those prompts cites what it can read and quote in the buyer's language, so multilingual answer coverage is usually the biggest and cheapest competitive gap a Thai business can close.
Which industries in Southeast Asia are most exposed?
High-consideration, research-heavy purchases, where buyers ask for a shortlist long before contacting anyone. The two we focus on: automotive (dealers, importers, EV service centres) and clinics serving medical tourism (cosmetic dentistry, hair transplant, surgery, LASIK, fertility). Both share the same shape: an expensive decision, heavy cross-border and cross-language research, and websites built for people rather than engines. Each has a dedicated page covering what AEO fixes in that vertical.
Group of Dots is an Answer Engine Optimisation studio. We get established businesses found, and chosen, when buyers ask AI for a shortlist.
Last updated: 17 July 2026
The question that matters is about your site.
Are AI engines citing you today? The AEO audit measures it page by page and returns the fixes attached.