Search is increasingly answered rather than listed. AEO is the work that decides whether that answer gets built from your page or from a competitor's — crawl access for AI agents, content an engine can lift a clean answer out of, and entities it can resolve without guessing.
Answer engines do not hand over ten links. They retrieve a handful of passages, synthesise one response and attribute a few sources. AEO is the work of getting your pages into that shortlist — retrievable, extractable, and consistent enough that a machine can quote you without guessing.
An answer engine — Google's AI Overviews and AI Mode, ChatGPT, Perplexity, Copilot, Claude — does not give the user a list to choose from. It fetches a set of passages, writes a single answer from them, and cites some of what it used. Three things decide whether you are in that set: whether the engine can fetch your page at all, whether it can lift a clean self-contained answer out of it, and whether what you say is corroborated by sources it already trusts.
It is not a replacement for SEO. Every answer engine still leans on a conventional index — Google's own, Bing's, or its own crawl. A page that classic search cannot crawl or index does not become visible to a language model because somebody added schema to it. AEO sits on top of technical SEO, not instead of it.
Evidence first, theory second. I run your real commercial questions through the engines your buyers actually use, record what comes back, and note who is being cited instead of you and why. That gives us a dated starting line we can re-run later — which matters, because none of these platforms give you a Search Console.
Answer engines extract passages, not pages. That rewards a particular shape: the question posed as a heading, answered directly in the first two or three sentences underneath it, in wording that still makes sense once it has been lifted away from the rest of the page. Most pages bury the answer under three paragraphs of warm-up. Restructuring what you already have usually beats writing something new.
Machines resolve entities, not keywords. Your name, business, services, locations and credentials need to be stated consistently on your own site and to match what third-party sources say about you. Organization, Person, Service, LocalBusiness and Article markup help a parser attach the right facts to the right entity. Schema will not get a weak page cited — it removes ambiguity from a good one.
AI crawlers are separate user agents — GPTBot, ClaudeBot, PerplexityBot, Google-Extended and others — and a surprising number of sites block them by accident through robots.txt, a CDN bot rule or a firewall setting nobody remembers enabling. I check what each agent is actually permitted to fetch, and confirm the page returns usable content without JavaScript, because most of them do not execute it.
Content written for citation reads differently: definitions stated plainly, figures attributed to a named source, comparisons set out in a table, no throat-clearing. I write the briefs and the pages themselves — mapped to questions people actually ask, structured for extraction, and edited so a human still wants to read it.
Local questions get answered from a different pool: your Google Business Profile, the wording of your reviews, directory consistency and your location pages. AEO for a local business is mostly local SEO done properly, plus content that answers what people ask in the days before they call.
There is no impressions report for AI answers. What can be measured: whether you are cited for a fixed set of tracked questions, re-tested on a schedule; referrals from AI platforms isolated as their own channel in GA4; and the branded-search and direct-traffic movement that tends to follow being mentioned in an answer nobody clicked.
Record the current answers for an agreed set of questions, with citations and dates. Without this you cannot tell later whether anything actually changed.
Pull them from Search Console queries, sales calls, your inbox and the related-questions surface — not from a keyword tool's guess at question volume.
robots.txt, CDN and firewall rules, and server-side rendering, so the agents can fetch what you publish.
Answer-first headings, self-contained passages, tables where a comparison is being made — and new pages only where nothing exists to restructure.
Entity and service schema, internal links that make the relationship between pages explicit, and consistency checks against your off-site profiles.
Re-run the baseline questions, show what moved, and say plainly what did not.
The terms overlap and the industry has not settled them. Working definitions that hold up: SEO is being found in a ranked list of links. AEO is being used in a direct answer — a featured snippet, an AI Overview, a chatbot reply. GEO, generative engine optimisation, is the subset of AEO aimed at systems that synthesise rather than quote verbatim. In practice one body of work serves all three, because a page that is crawlable, clearly structured and factually consistent does well in every one of them. Be sceptical of anyone selling them to you as three separate retainers.
Answer engine optimisation is the work of getting a page used and cited when a search engine or AI assistant answers a question directly instead of returning a list of links. It covers crawl access for AI agents, content structured so a clean answer can be lifted out of it, consistent entity and schema markup, and tracking whether you are actually being cited.
It is an extension of SEO, not an alternative to it. Answer engines draw on conventional indexes, so a page that cannot be crawled or indexed cannot be cited. What AEO adds is attention to passage-level extractability and entity clarity rather than to ranking position.
Not for Google rich results. Google stopped showing FAQ rich results on 7 May 2026 and withdrew the reporting for them shortly afterwards. FAQPage markup is still valid Schema.org and it does no harm, but leaving it in place is not a strategy. What still works is the FAQ content itself — a clear question as a heading with a direct answer underneath, which is exactly the shape an answer engine extracts.
Almost certainly not. It is a proposed convention that the major engines have not committed to using, and Google has said publicly that it does not use it. Publishing one costs very little, but anyone selling it as an AEO deliverable is selling you a file rather than a result. Your robots.txt, your rendering and your content structure are what actually decide whether an AI system can use your page.
Some of it is visible — referrals from AI platforms show up in GA4 and can be isolated into their own channel. Much of it is not: a cited answer the user never clicks leaves no referral at all, and often surfaces later as a branded search or a direct visit. That is why the tracked-question baseline matters. It measures citation directly instead of inferring it from traffic.
Access and structural fixes can change what an engine retrieves within a few weeks of re-crawling. Entity consistency and citation for competitive questions take longer, because they depend on corroboration from sources you do not control. I would judge a first engagement at the three-month mark, against the baseline recorded at the start.
No, and nobody can. The answer set is generated per query, varies between users and changes without notice. What I can do is make sure you are retrievable, extractable and consistent — and show you, with dated evidence, what changed after the work.
Both. Most AEO gains come from restructuring pages you already have, which I do directly. Where a question has no page to answer it, I write the page — briefed from real queries and edited to be worth reading rather than padded to a word count.
Send me your domain and the three or four questions a good customer would ask before buying. I will run them through the answer engines, tell you what comes back and who is being cited — and whether the gap is a technical one, a content one, or nothing worth spending money on yet. Start a conversation here.
“Fantastic work. Kevin will 100% be working again soon!”
“John Kevin has been great. He is well versed in SEO content deliverables and is a very fast learner. We appreciate everything he has been able to do for us.”
“Excellent experience working with Kevin. Very organized, fast learner, meticulous and detail oriented. Will hope to hire again in the future.”
Tell me what you are trying to improve and I will tell you honestly where I can help — and what I would look at first. No pitch deck, no 40-slide audit you will never read.
Based in the Philippines, working with clients across the US, Europe and Asia-Pacific. I reply to every enquiry personally.