--- title: "The New Battlefield for Online Visibility Is Called Answer Engine Optimisation (AEO)" date: 2026-03-15 description: "Discover what Answer Engine Optimisation (AEO) is and how to structure your content to be cited by ChatGPT, Perplexity, and Google AI. A practical guide with real-world examples, Schema Markup, and a final checklist." author: "Petru Cojocaru" categories: - name: "SEO optimization" url: "https://cojocarupetru.info/seo-optimization.md" --- # The New Battlefield for Online Visibility Is Called Answer Engine Optimisation (AEO) **And if you haven't heard of AEO, you're already losing ground.** I noticed the first sign on a Tuesday morning, when I opened Analytics as usual. The traffic was there — impressions were rising, positions remained stable, older articles were holding their places in the top 5. On paper, everything looked fine. But something didn't add up: actual visits had been declining slightly, steadily, for several months. Not dramatically. Not catastrophically. Just a thin thread slowly unravelling from a ball I thought I was holding firmly in my hand. I began looking for technical explanations. Loading speed, Core Web Vitals, indexing issues. Everything was in order. And yet, something had changed. Not in the site. In the way people search. One evening, I ran a simple experiment. I took 10 questions from my field — questions for which I knew I had well-written, well-ranked articles. I put them one by one into ChatGPT. Then into Perplexity. Then I asked Google Assistant. The answers came immediately, complete, with sources. My site didn't appear in any of them. Others did — smaller sites, with less authority, with less traffic. But structured differently. Written differently. That was when I understood I wasn't losing an SEO battle. I was losing a battle I didn't even know I was playing. It's called **Answer Engine Optimisation** — AEO. And it is, at least for me, the most significant change in online visibility over the last 10 years. It is not a term invented by a marketing agency to sell new services. It is a measurable reality: according to Sparktoro data, nearly 60% of Google searches today end without a single click. The user receives the answer directly — on screen, by voice, through AI — and leaves. Without visiting anyone. Without leaving a trace in Analytics. I spent several months studying how this new world of search functions. What follows is everything I learned — structured, with real examples, with the mistakes I made, and with what ultimately worked. --- ## 1. What AEO Is and Why It Differs from Everything We Knew About SEO ### What Is the Real Difference Between SEO and AEO? Classic SEO has a simple logic I internalised over years of practice: you optimise the page, build links, climb the rankings, earn clicks. The currency is the visit — the user lands on your site and everything begins from there. AEO changes this logic fundamentally. The goal is no longer to bring the person to your site. The goal is for your answer to reach the person — directly, via an AI engine, without them having to make the journey to you. **SEO = you want to be found. AEO = you want to be cited.** It seems like a small difference. It isn't. A site optimised solely for SEO can rank first on Google and be completely absent from AI responses. A smaller blog, with correctly structured content, can be cited daily by ChatGPT or Perplexity without appearing in the top 5 organic results. **An example from my own practice:** I have an old article about password security — written in 2021, well documented, over 2,000 words long. It ranked second for several relevant queries. When I tested it on Perplexity, the article didn't appear as a source at all. I investigated why: the introduction took up 300 words without a single direct answer. The section headings were literary, not interrogative. There was no question-and-answer structure. I rewrote the article in AEO format — H3 headings as questions, a direct answer within the first 50 words after each heading, FAQPage Schema in the code. Within 3 weeks, Perplexity was citing it for 4 different queries about password security. Direct clicks didn't increase dramatically. But presence in AI responses — that did. ### How Has the Rise of AI Changed the Way People Search for Information? I remember how I used to search for information 10 years ago. I'd open Google, scan the first 3 or 4 results, open 2 or 3 tabs, read, compare, synthesise for myself. It was an active process — almost heuristic. Today, I do something different. I ask ChatGPT. I receive a structured response, with nuance, with sources. If I need more, I ask for more. I visit an external site only if I have a specific reason — to verify a source, to buy something, to contact someone. My behaviour has changed. And, more importantly, the behaviour of my readers has changed in the same way — even if they don't necessarily realise why. ### Why 'Position Zero' Is Only a First Signal, Not the Destination The featured snippet — the block of text that appears above all Google results — was for a long time the supreme trophy of SEO. And it remains valuable. But it is an indicator of a broader ecosystem, not an end in itself. Today, real visibility means being cited in ChatGPT, appearing as a source in Perplexity, being the answer Google Assistant gives when someone asks a question aloud. The featured snippet is the first step on a longer road. AEO is the entire journey. ### What Types of Query Are Affected Differently by AEO? Not all searches behave the same way in the AI era, and treating AEO as a single unified strategy is one of the mistakes I observed frequently — including in my own early months of experimentation. There are at least three distinct categories, each with its own dynamics: - **Simple factual queries** ('who founded company X', 'when did the term Y first appear') are today dominated by AI Overviews and responses generated from training data. Citation of external sources is rare or superficial. Optimising for this category builds brand visibility, not necessarily traffic. - **Complex, in-depth queries** ('how do you build a content strategy for a niche market', 'why do digital transformation projects fail') remain the area where authoritative sources are cited and accessed. The user posing such questions is seeking not just an answer, but also a voice, a context, a perspective. These are the queries for which it is genuinely worth writing. - **Commercial queries** ('best WordPress hosting', 'VPN comparison 2025') remain highly competitive for traditional SEO. Purchase intent continues to attract clicks, even in the presence of AI Overviews. Knowing which category your subject falls into changes your optimisation strategy entirely. --- ## 2. How Answer Engines Think — and Why It Matters for How You Write ### How Do AI Models 'Read' a Blog Article? This was, for me, one of the central revelations. I had long assumed that AI models read an article the way a well-read person does — appreciating narrative structure, the flow of ideas, stylistic subtleties. That is not the case at all. AI platforms do not operate uniformly, and it is important to understand the differences between them before building your strategy: - **Perplexity** uses a RAG (Retrieval-Augmented Generation) architecture with real-time indexing and transparent citation. Your content can be accessed and cited even if it was published recently. Clear structure and direct answers matter enormously here. - **ChatGPT Search** and **Claude** have indexing and source-selection methods that are not publicly documented. They operate partly from trained knowledge, partly from real-time web access — the proportion and criteria are not disclosed. - **Google AI Overviews** is closest to traditional SEO: domain authority, backlink signals, and Schema Markup all play a demonstrable role. One useful practice — though sometimes oversimplified in the specialist literature — is the rule of a 'direct answer within the first 50–60 words after the heading'. This works well for Perplexity and Google featured snippets. But not all models prioritise exclusively this textual window — the semantic context of the whole article and the authority of the author entity matter just as much for some platforms. Clear structure remains essential; but it is not a magic formula applicable identically everywhere. Everything I ever wrote with elaborate introductions, with context built gradually, with the answer revealed only halfway through the article — these are harder for an AI answer engine to extract. Not because they are poor. But because they do not follow the architecture that AI prioritises. ### What Role Does NLP Play in the Way AI Understands My Content? NLP — Natural Language Processing — is the technology that enables an AI engine to understand the intent behind a search, not just the exact words. If someone writes 'how do I protect my site from hackers' and another writes 'website cybersecurity', NLP understands that both want the same thing. For me, this changed the way I think about keywords. I no longer optimise for an exact phrase. I optimise for an intent, a question, a life context of my reader. I write for people — but I structure for algorithms. ### How Does a Vocal Answer Engine Choose the Source It Cites? When I first tested Google Assistant with questions from my field, I was struck by the consistency of its selection. It wasn't necessarily the large sites that won. It was the sites with clear, concise, structured answers — regardless of size. The main criteria are: domain authority, loading speed, HTTPS, and — most importantly for us — a direct answer available immediately after the heading, in short sentences, without ambiguity. --- ## 3. How I Restructured My Content for AEO ### How Do I Write H2 and H3 Headings Now? I completely changed the way I formulate section headings. Previously I wrote editorial headings — suggestive, sometimes poetic, always designed to capture human attention. 'The Anatomy of a Lost Click.' 'When Google Is No Longer the Answer.' I liked them. They looked good. The problem: no AI engine could make easy use of them. They contained no question. They gave no signal that a direct answer was to follow. Now I write complete interrogative headings — the exact phrasing a real reader would type or speak aloud: **Before:** 'The Anatomy of a Lost Click' **After:** 'Why Is Organic Traffic Declining Even Though Google Rankings Remain Stable?' The content within the section can remain just as nuanced, just as personal. But the heading must be the question. Always. ### What Is the Inverted FAQ Format and How Do I Use It in Practice? I discovered the inverted FAQ format after reading several analyses of the structural patterns of pages frequently cited by AI. The logic is counterintuitive for a classically trained writer: - **The H3 heading** = the complete question, formulated as a genuine question - **The first paragraph** = the direct answer, 40–60 words, with no introduction, no prior context - **The rest of the section** = nuance, context, personal experience, examples As a writer, it took me some time to accept that I had to give the answer before building the context. The narrative instinct says the opposite — you build the tension, then release the answer. AEO says: give the answer first. The story comes after. ### How Do I Create a 'Quick Answer' Box at the Beginning of Articles? A simple technique I've adopted for nearly all new articles: a highlighted block immediately after the introduction, with the following structure: > **Quick answer:** > > > - Essential point 1 — in a maximum of 20 words > - Essential point 2 — in a maximum of 20 words > - Essential point 3 — in a maximum of 20 words This block does two things simultaneously: it helps the reader in a hurry decide whether the article is useful to them, and it offers the AI exactly the condensed format it seeks in order to generate a structured response. ### How I Learnt to Write Shorter Sentences Without Losing My Personal Voice This was the most difficult adaptation. I write naturally in long sentences — with subordinate clauses, with digressions, with parenthetical asides. I like the way they sound. I build my rhythm from them. But for AEO, the practical rule is clear: sentences of no more than 20 words, paragraphs of no more than 4 lines, where we are aiming for AI extraction. Anything longer risks being passed over by answer engines. The solution I found: I write the first draft the way I always write — with my voice, with my rhythm. Then I take a second, separate pass, in which I cut and restructure for clarity. The voice remains. The length is reduced. **An example from my own practice:** I rewrote an article about digital identity that had been written entirely in my narrative style — dense paragraphs, sentences of 35–40 words, no interrogative structure whatsoever. I kept the substance almost identical, but restructured each section: H3 heading as a question, a direct 50-word answer opened immediately, the rest of the narrative afterwards. Long sentences cut in two where the meaning permitted. I added FAQPage Schema. Within 30 days, the article appeared as a cited source in Perplexity for 3 queries about online identity and digital presence. The blog comments hadn't changed — the voice had remained the same. But visibility in AI had increased significantly. --- ## 4. Technical Implementation: Schema Markup — a Useful Tool, Not a Magic Solution Schema Markup was for a long time the area I kept putting off. It seemed technical, it seemed to require coding knowledge, it seemed to be for large sites with dedicated teams. I was wrong on all three counts. But I was also wrong in the other direction, after I discovered it: I overestimated it. Schema does not guarantee visibility in AI and is not a universal shortcut. It is a useful signal — more so for some platforms than for others. ### What Type of Schema Markup Is Essential for a Personal Blog? There are four types I use or recommend for an editorial-style blog: - **FAQPage Schema** — for any section with a question-and-answer structure. Most effective for Google AI Overviews and featured snippets, relatively straightforward to implement. - **Article Schema** — for standard blog posts. It tells engines that the page is an editorial article with an author, a date, and a subject. - **HowTo Schema** — for practical guides with clear steps. If you write tutorials or instructions, this is your format. - **Person Schema** — often forgotten, but crucial for a personal blog. It links content to your identity as an author and contributes to building your digital entity. ### What Are the Limitations of Schema Markup in the Context of AI? A Search/Atlas study published in December 2024 found no direct, systematic correlation between Schema coverage and citation rate in generative AI responses. This conclusion deserves to be taken seriously. Schema demonstrably assists platforms that explicitly acknowledge it — Google AI Overviews and Bing Copilot. There is no public evidence that ChatGPT or Perplexity use it systematically as a source-selection signal. Perplexity prioritises content relevance and RAG indexing accessibility; ChatGPT has undisclosed criteria. My practical conclusion: implement Schema because it helps Google and because it is good structural practice. Do not treat it as the primary solution for citations in ChatGPT or Perplexity. ### How Do You Validate Schema Without Advanced Technical Knowledge? Two free tools I use constantly: - **Google Rich Results Test** (search.google.com/test/rich-results) — enter a URL and see exactly what Google understands from the structure of your page. Simple, visual, fast. - **Schema Markup Validator** (validator.schema.org) — more technical, more detailed. Useful when you want to understand precisely where an implementation error lies. On WordPress, the Yoast SEO or RankMath plugins generate Schema automatically for Article and FAQPage. There is no manual coding involved. It's a ticked option and a completed field. ### What Is the Connection Between Schema and Voice Search Responses? Direct and demonstrable — within the Google ecosystem. When Google Assistant receives a spoken question, it looks for pages with correctly implemented Schema as a priority. It cannot guess the structure of free-form text — it needs explicit signals from the code. FAQPage Schema is, in my experience, the shortest route to Google voice responses and featured snippets. If you have an FAQ section on an important page and you add the correct Schema, your chances of being the voice answer for those questions increase significantly. **An example from my own practice:** On one of the blog's resource pages, I added FAQPage Schema with 8 frequently asked questions about writing online and building a blog in Romania. The implementation took less than an hour with a plugin. Within two weeks, 3 of the 8 questions appeared as featured snippets on Google. One of them became the Google Assistant answer to the corresponding voice query. I hadn't changed the content. I had simply added the language that Google's machines understand. --- ## 5. How I Write Direct Answers That AI Cites ### What Is the Ideal Length of an Answer to Be Picked Up by an AI Engine? I tested several lengths and formats systematically. My practical conclusion: - **Answer for a text featured snippet**: 40–60 words. Complete enough to be useful, short enough to be reproduced in full. - **Voice answer**: 1–2 sentences, a maximum of 25–30 words. This is the format someone hears when they ask a question aloud. The strategy I use: I write both variants in the same place. The first sentence or two — the extremely condensed answer, for voice. The full 40–60-word paragraph — the standard answer, for text. ### How Do I Use Concrete Data to Earn the AI's Trust? AI models prefer sources that back up their claims with verifiable figures. Not because they 'know' the data is correct — but because they have learnt that credible sources anchor assertions in measurable reality. I changed an old habit: where I used to write 'many people prefer', I now write 'according to a Nielsen study from 2024, 67% of users prefer'. Where I used to write 'slow sites lose visitors', I now write 'Google reports that 53% of mobile users abandon a page that takes more than 3 seconds to load'. That doesn't mean the personal style disappears. It means that important claims are anchored. ### How Do I Answer Complex 'Why' and 'How' Questions? These are the questions I love as a writer — they have depth, they admit nuance, they allow exploration. And they are, at the same time, the most valuable for AEO, because they are precisely the type of conversational queries people pose to AI assistants. The structure that works best, without sacrificing editorial quality: - **The direct answer in the first sentence** — without preamble, without prior context - **A list of reasons or steps** — in bullet-point or numbered format - **The story or personal example** — this is where my voice and experience come in - **A short conclusion** — a maximum of 2 sentences The direct answer satisfies the AI. The personal story satisfies the human reader. You don't have to choose between them. **An example from my own practice:** I published an article about why Romanian bloggers give up after their first year. The heading was interrogative, the structure was clear. The first paragraph gave a direct answer in 45 words, followed by 5 H3 sections, each with a direct answer and a personal story. I supported each point with a statistic or study. The article appeared as a cited source in Perplexity for queries about blogging in Romania. It received organic links from two digital marketing platforms. Comments — the ones that truly matter for a personal blog — increased rather than decreased. AEO restructuring didn't impoverish the article. It made it clearer for both person and machine. --- ## 6. Case Study: How I Transformed an Old Article into One Cited by AI ### Before vs. After: a Real Article Rewritten I chose for the experiment one of the blog's oldest articles — a text about building an authentic digital presence. Written in 2020, with a personal voice, with stories, with reflections. I valued it. And yet it never appeared in AI responses. **The original version began like this:** > *'There is a paradox I observe in many people who try to build an online presence: the harder they try to appear authentic, the more artificial they become. As though authenticity were a suit one puts on in the morning...'* Beautiful? Perhaps. Useful for an answer engine? Completely invisible. **The AEO version of the same section:** > **How Do You Build an Authentic Digital Presence in 2025?** > > **Quick answer:** An authentic digital presence is built through consistency of voice, not visual perfection. Publish regularly, write about real experiences, respond to comments. Online authenticity is not declared — it is demonstrated over time. The story about the paradox of the suit didn't disappear. It moved down into the details section, where it belonged. The direct answer came first. ### What Does a Correctly Structured AEO Article Look Like in Code? At a technical level, the structure of an AEO article includes: - `

` and `

` tags containing the complete questions - A `

` paragraph immediately after each heading, with the direct answer in the first 50 words - `