{"id":8476,"date":"2026-08-07T09:33:00","date_gmt":"2026-08-07T09:33:00","guid":{"rendered":"https:\/\/www.voctos.com\/?p=8476"},"modified":"2026-09-21T11:41:23","modified_gmt":"2026-09-21T11:41:23","slug":"ask-engine-optimization","status":"publish","type":"post","link":"https:\/\/www.voctos.com\/ar\/blog\/ask-engine-optimization\/","title":{"rendered":"\u062a\u062d\u0633\u064a\u0646 \u0645\u062d\u0631\u0643\u0627\u062a \u0627\u0644\u0625\u062c\u0627\u0628\u0629 (AEO) \u0641\u064a 2026: \u0645\u0627 \u0627\u0644\u0630\u064a \u064a\u0646\u062c\u062d\u060c \u0648\u0645\u0627 \u0627\u0644\u0630\u064a \u064a\u062b\u0628\u062a \u0623\u0646\u0647 \u0644\u0627 \u064a\u0646\u062c\u062d"},"content":{"rendered":"\n
VOCTOS<\/a> › The GEO Guide<\/a> › This guide<\/span><\/div>\n\n\n\n
Last updated 20 August 2026  \u00b7  Every claim linked to a primary source  \u00b7  Includes an Arabic and Gulf section no competing guide covers<\/div>\n\n\n\n
−4.6%<\/div>
the measured effect of adding schema on AI Overview citations, with controls<\/div><\/div>
97%<\/div>
of llms.txt files received zero requests in a month<\/div><\/div>
0.04<\/div>
correlation between page word count and being cited<\/div><\/div>
2014<\/div>
the year the voice search statistic everyone quotes was actually said, by someone else, about something else<\/div><\/div>
68%<\/div>
of US Google searches now end without a click<\/div><\/div><\/div>\n\n\n\n

Search “ask engine optimization” and you will find agency pages telling you to write FAQ pages, add schema markup, optimize for voice search, and keep your answers between forty and sixty words. Almost none of that survives contact with the published evidence, and the top-ranking page for the term does not implement the schema it tells you to implement.<\/p>\n\n\n\n

This guide takes the opposite approach. Every claim below is linked to a primary source, including the ones that argue against things we sell. Where the evidence is genuinely uncertain, it says so.<\/p>\n\n\n\n

FIRST, THE TERM ITSELF<\/div>
“Ask Engine Optimization” is a coined agency variant. The established expansion is Answer<\/strong> Engine Optimization, and it is the one Google uses in its own documentation. Insivia, the agency that popularized “Ask”, has since added an editor’s note to its own page deprecating the term in favor of “Answer”. The phrase appears nowhere in Google Search Central, Search Engine Land, Search Engine Journal, the Ahrefs blog, the Semrush blog or Wikipedia. We use it in the title because that is what people search for. We will not pretend it is standard.<\/div><\/div>\n\n\n\n

Key takeaways<\/h2>\n\n\n\n
1<\/div>
It is Answer, not Ask<\/div>
Google’s own documentation names AEO and GEO. The agency that popularized “Ask” publicly retired the term.<\/div><\/div><\/div>
2<\/div>
Schema does not move AI citations<\/div>
A controlled study of 1,885 pages that added JSON-LD found −4.6% on AI Overviews and effects indistinguishable from zero on AI Mode and ChatGPT. Google states outright that structured data is not required.<\/div><\/div><\/div>
3<\/div>
llms.txt does essentially nothing<\/div>
Across 137,210 domains, 97% of llms.txt files received zero requests in a month. There were zero AI bot requests for files that do not exist.<\/div><\/div><\/div>
4<\/div>
The voice search advice is a decade-old misquote<\/div>
“50% of searches will be voice by 2020” came from Baidu’s chief scientist in 2014, referred to images and speech, and was about Baidu. comScore has explicitly denied publishing it.<\/div><\/div><\/div>
5<\/div>
Formatting tweaks are not the lever<\/div>
A study of 252,000 controlled trials across six models found topical relevance and list position drive citation, and that formatting-only edits have little impact.<\/div><\/div><\/div>
6<\/div>
The strongest correlate is being talked about elsewhere<\/div>
Across 75,000 brands, YouTube mentions correlate with AI visibility at 0.737. Publishing more pages correlates at 0.194.<\/div><\/div><\/div><\/div>\n\n\n\n

What an answer engine actually does before it quotes you<\/h2>\n\n\n\n

Most guides skip the mechanism and go straight to tactics, which is why the tactics are usually wrong. Google has documented the process in its own words.<\/p>\n\n\n\n

When you ask a question, the engine does not run one search. It performs query fan-out<\/strong>breaking your question into subtopics and issuing a multitude of queries simultaneously<\/a>. Google’s own worked example: “how to fix a lawn that’s full of weeds” becomes “best herbicides for lawns”, “remove weeds without chemicals” and “how to prevent weeds in lawn”. Deep Search can issue hundreds of these.<\/p>\n\n\n\n

It then grounds<\/strong> the answer by relying on core Search ranking systems to retrieve relevant, up-to-date pages from the Search index<\/a>. The gate is simple and worth memorizing: a page must be indexed and eligible to be shown in Google Search with a snippet<\/a>. If you block snippets, you are not eligible.<\/p>\n\n\n\n

THE PRACTICAL CONSEQUENCE OF FAN-OUT<\/div>
The query you type is not the query that retrieves anything. That has two implications. First, optimizing a page for the literal question is optimizing for the wrong key. Second, and this is where a lot of agencies go wrong, Google explicitly states that creating separate content for every fan-out variation primarily to manipulate generative responses violates its scaled content abuse spam policy<\/a>. Cover the subtopics within a page. Do not build a page per subtopic.<\/div><\/div>\n\n\n\n
TAKEAWAY<\/strong>
Answer engines break a question into many parallel sub-queries and then retrieve against those, not against the question as typed. Being indexed and snippet-eligible is the entry gate; everything else is a refinement on top of it.<\/div><\/div>\n\n\n\n
<\/th>SEO<\/th>AEO (Answer Engine Optimization)<\/th>GEO (Generative Engine Optimization)<\/th><\/tr><\/thead>
What it optimizes<\/td>A page, to be clicked<\/td>A passage, to be quoted as the answer<\/td>A brand entity, to be synthesized into a generated response<\/td><\/tr>
Where the answer appears<\/td>On your site, after a click<\/td>In the results page itself, featured snippets, People Also Ask, AI Overviews<\/td>Inside ChatGPT, Perplexity, Gemini, Copilot<\/td><\/tr>
Primary success signal<\/td>Position and click-through rate<\/td>Being the extracted answer<\/td>Being mentioned and described accurately, with or without a link<\/td><\/tr>
Does the user visit you?<\/td>Yes, that is the point<\/td>Often not<\/td>Usually not: about 19% click through to a cited source<\/td><\/tr>
Is it a separate discipline?<\/td><\/td>Google’s position: no, it is still SEO<\/td>Google’s position: no, it is still SEO. Practitioners disagree on the measurement layer, not the fundamentals<\/td><\/tr>
What genuinely differs<\/td><\/td>Answer placement and structure within the page<\/td>Sources off your own site, and engine-specific, non-deterministic measurement<\/td><\/tr><\/tbody><\/table>
The honest version. The three overlap heavily, and Google states in its own documentation that optimizing for generative AI search is still SEO.<\/figcaption><\/figure>\n\n\n\n

What the evidence says actually drives citation<\/h2>\n\n\n\n

The most rigorous experiment published on this ran 252,000 controlled trials<\/strong> across six large language models and eighteen content factors, varying one factor per trial, anonymizing brands and counterbalancing order. Its conclusion: topical relevance and list position are the biggest drivers<\/a>, explicit price information and a recent timestamp also help, and formatting-only edits have little impact<\/strong>.<\/p>\n\n\n\n

That last clause is the one to sit with, because formatting is what most AEO advice consists of.<\/p>\n\n\n\n

\"Bar
Spearman correlations with AI brand visibility across 75,000 brands. Source: Ahrefs, 12 December 2025.<\/figcaption><\/figure>\n\n\n\n

A separate analysis of 75,000 brands found the strongest correlate of AI visibility is YouTube mentions<\/strong>, at a Spearman correlation of 0.737, ahead of branded web mentions (0.66–0.71), branded search volume (0.35–0.47) and Domain Rating (0.27–0.33). The number of pages on your site correlates at 0.194, which the authors describe as almost no relationship.<\/p>\n\n\n\n

And on the length question that every FAQ guide answers confidently: Ahrefs analyzed 174,048 pages cited across 560,346 AI Overviews and found the correlation between word count and citation is 0.04, essentially zero. The average cited page runs 1,282 words, but 53.4% of cited pages are under 1,000 words and 16.6% are under 350. More than 95% of citations to very short content land in the top three citation positions.<\/p>\n\n\n\n

TAKEAWAY<\/strong>
Across 252,000 controlled trials, topical relevance and list position drove citation while formatting-only edits had little effect. Page length shows essentially no correlation with being cited, which makes most word-count prescriptions unfounded.<\/div><\/div>\n\n\n\n

Three things you have been told to do that the evidence does not support<\/h2>\n\n\n\n

1. Schema markup<\/h3>\n\n\n\n
\"Chart
The apparent schema effect before and after matched controls. Source: Ahrefs, 11 May 2026.<\/figcaption><\/figure>\n\n\n\n

The correlation is real and it is misleading. AI-cited pages are almost three times more likely to carry JSON-LD, and 53% of AI-cited pages run schema. That is the statistic every AEO guide quotes.<\/p>\n\n\n\n

Then someone ran the causal test. Ahrefs took 1,885 pages that added JSON-LD between August 2025 and March 2026 and compared them against roughly 4,000 matched controls. Raw before-and-after growth on AI Mode was +43%. After subtracting the controls it was +2.4%, statistically indistinguishable from zero. ChatGPT came in at +2.2%, also indistinguishable. Google AI Overviews came in at −4.6%<\/strong>, and that one was statistically significant.<\/p>\n\n\n\n

Their explanation is the important part: schema markup tends to live on better-maintained, more technically sophisticated sites. The correlation was measuring the site, not the markup.<\/p>\n\n\n\n

A separate experiment tested whether AI systems can even read it. Researchers exposed a price only in JSON-LD and asked five systems to retrieve it. Zero of five extracted it<\/a>. Hidden Microdata and hidden RDFa also scored zero of five.<\/p>\n\n\n\n

And Google states it plainly: “Structured data isn’t required for generative AI search, and there’s no special schema.org markup you need to add.”<\/a> It then recommends continuing to use structured data anyway, for rich results, which is a different and still-valuable job.<\/p>\n\n\n\n

SO SHOULD YOU STILL USE SCHEMA?<\/div>
Yes, and for honest reasons. It earns rich results in classic search, it makes commerce data machine-readable in a way prose cannot, and John Mueller has noted that pricing, shipping and availability are near-impossible to read accurately from a text page. What you should stop doing is selling schema as an AI citation tactic, or expecting it to move AI visibility on its own. Ahrefs also flags a genuine limit on its own study: every page it tested was already being cited, so the study cannot speak to whether schema helps a page get seen in the first place.<\/div><\/div>\n\n\n\n

2. llms.txt<\/h3>\n\n\n\n
\"Donut
What actually happened to llms.txt files across 137,210 domains. Source: Ahrefs, 15 June 2026.<\/figcaption><\/figure>\n\n\n\n

Ahrefs checked 137,210 domains with traffic in May 2026. About 28% published an llms.txt file. 97% of those files received zero requests that month.<\/strong> Of the 3% that got any traffic, 96% was bots and the single largest category was SEO audit tools at 21.7%, followed by the SEO and GEO industry itself studying the file at 12.1%. Actual AI retrieval bots, OAI-SearchBot, PerplexityBot, Claude’s search crawler, accounted for 1.1%.<\/p>\n\n\n\n

The finding that ends the debate: there were zero AI bot requests for llms.txt files that do not exist<\/strong>. The 404 probes were 98% human. AI systems do not go hunting for the file; they fetch it when a link, an index or a user instruction tells them it is there. Ahrefs also notes that Slackbot fetched llms.txt files more often than PerplexityBot did.<\/p>\n\n\n\n

Google is unambiguous: “Google Search itself doesn’t use them… Google Search ignores them.”<\/a> No vendor, not OpenAI, not Anthropic, not Perplexity, documents its crawler reading yours, even though several publish one of their own.<\/p>\n\n\n\n

THE ONE NARROW CASE WHERE IT IS DEFENSIBLE<\/div>
AI coding agents. In the same study, agents were the largest AI category at 10.5% and Claude-Code was the second-biggest AI fetcher. If you publish developer documentation and want coding agents to consume it efficiently, llms.txt is a reasonable developer-experience decision, which is precisely what it was proposed for in September 2024. It is not a search visibility deliverable, and it should not appear on an AEO invoice.<\/div><\/div>\n\n\n\n

3. Voice search optimization<\/h3>\n\n\n\n

The statistic underneath a decade of voice search advice is misattributed. “Fifty percent of all searches will be voice by 2020” was said by Andrew Ng, then chief scientist at Baidu, in September 2014<\/a>and what he actually said was “images or speech”, about Baidu, in China. comScore, to whom it is routinely credited, told a researcher in writing that it never published the prediction<\/a>.<\/p>\n\n\n\n

Google’s only official voice figure is Sundar Pichai’s May 2016 statement that 20% of queries on the Google mobile app and Android were voice<\/a>. It has not been updated in a decade. The modern “20.5% use voice search” figure is a different metric, share of internet users, not of queries, and it peaked at 22.5% in 2022 and has declined since.<\/p>\n\n\n\n

The measurement infrastructure has been dismantled. Edison Research’s Infinite Dial 2026<\/a> dropped voice assistant usage entirely and replaced it with an AI usage module. In that same study, 57% of Americans aged 12 and over have used a generative AI assistant. Pew found 49% of US adults now use AI chatbots and 42% use them to search for information<\/a>.<\/p>\n\n\n\n

The assistants themselves have been rebuilt as answer engines. Alexa+ runs on Amazon Nova and Anthropic models; Apple’s next-generation Siri goes to the web and generates an answer. Across four official Amazon and Apple announcements, neither company mentions source citations, publisher attribution or clickable links.<\/p>\n\n\n\n

TAKEAWAY<\/strong>
Schema markup, llms.txt and voice search optimization are the three most commonly sold answer-engine tactics, and each is either unsupported by controlled evidence or explicitly disclaimed by the platform it targets. Two of them remain worth doing for other reasons.<\/div><\/div>\n\n\n\n

What does work, ranked by strength of evidence<\/h2>\n\n\n\n
What to do<\/th>Why it is on this list<\/th>Evidence strength<\/th><\/tr><\/thead>
Be indexed and snippet-eligible<\/strong><\/td>Google states a page must be indexed and eligible to be shown with a snippet to appear in AI features. This is the entry gate, not an optimization.<\/td>Primary<\/strong>Google documentation<\/td><\/tr>
Rank well for the topic, not just the query<\/strong><\/td>Citation probability tracks organic rank: 43% at position one, 31% at position three, 7% at position twenty. Ranking is no longer the gate it was, but it is still the strongest lever you control.<\/td>Strong<\/strong>362,000 queries, 5.1M citations<\/td><\/tr>
Get talked about on YouTube and across the web<\/strong><\/td>YouTube mentions correlate with AI visibility at 0.737, branded web mentions at 0.66–0.71, far ahead of anything on your own site.<\/td>Correlational, 75,000 brands<\/td><\/tr>
Publish ranked comparison content<\/strong><\/td>The ranked best-of listicle is the single most-cited content format, at roughly 21–22% of all AI citations, and is cited 2.7 times more often for informational queries.<\/td>Two independent datasets agree<\/td><\/tr>
Put the answer in the first section<\/strong><\/td>44.2% of AI citations are extracted from a page’s introduction or first section. This is placement, not formatting.<\/td>Moderate, 25,000 URLs<\/td><\/tr>
Add citations, quotations and statistics<\/strong><\/td>The foundational GEO paper measured 30–40% improvements from these. Read the caveat below before you act on it.<\/td>Primary<\/strong>but conditional<\/td><\/tr>
Keep a visible, honest last-updated date<\/strong><\/td>Recency is one of the factors the 252,000-trial study found does help, and pages with a clear timestamp are reported to receive more citations.<\/td>Moderate<\/td><\/tr>
Improve fluency and readability<\/strong><\/td>The GEO paper measured a 15–30% visibility boost from stylistic clarity alone.<\/td>Primary<\/strong><\/td><\/tr><\/tbody><\/table>
Ordered by how much confidence the underlying evidence actually supports, not by how often the tactic is sold.<\/figcaption><\/figure>\n\n\n\n
THE CAVEAT NOBODY REPORTS ABOUT THE 40% FIGURE<\/div>
The famous GEO paper result is rank-dependent, and it reverses<\/strong>. For sources appearing at rank five, adding citations improved visibility by 115.1%, quotations by 99.7% and statistics by 97.9%. For sources already at rank one, the same three changes measured −30.3%, −22.9% and −20.6%<\/strong>. Lower-ranked sources benefit; already-dominant ones can be harmed. The same paper found keyword stuffing produced little to no improvement and an authoritative tone produced no significant improvement at all. If you are already the cited source, do not “optimize” the page that is working.<\/div><\/div>\n\n\n\n

There is also a field-level verdict worth knowing before you buy any of this from anyone. A survey of 45 studies published between November 2023 and July 2026 concluded that no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability<\/a>, that the GEO paper’s gains are conditional on a source already being present in a fixed context, and that citation-oriented rewrites can impair retrieval.<\/p>\n\n\n\n

TAKEAWAY<\/strong>
The most-quoted result in this field is conditional on rank: lower-ranked sources gained up to 115% from adding citations and quotations, while sources already ranked first lost 20 to 30% from the identical change.<\/div><\/div>\n\n\n\n

The traffic reality, stated honestly<\/h2>\n\n\n\n

Two things are true at once, and most articles only tell you one of them.<\/p>\n\n\n\n

Zero-click search has accelerated.<\/strong> 68.01% of US Google searches ended without a click between January and April 2026<\/a>, up from 60.45% in 2024. Pew found that when an AI summary appears, 8% of visits produced a click on a traditional result versus 15% without one<\/a>, and 1% clicked a link inside the summary itself. Click-through rate at position one is down 58% where an AI Overview appears.<\/p>\n\n\n\n

But AI referral traffic is still tiny.<\/strong> Semrush measured AI at 0.14% of all web visits across 50,000+ sites in 2025, against organic search at 16.04%. Google AI Mode was 0.01%. SparkToro found AI Mode accounts for 0.34% of searches and noted plainly that if you theorised AI Mode caused the zero-click acceleration, it is not there yet.<\/p>\n\n\n\n

AND THE CONVERSION CLAIMS DO NOT HOLD UP<\/div>
You will see “AI traffic converts 4.4 times better” and “23 times better” quoted constantly. The 23-times figure comes from a single website over 30 days. Eight days after publishing it, the same company published a study of 81,947 sites and concluded: “is AI traffic really better quality? Not quite.” The largest peer-reviewed test, 973 e-commerce sites, $20 billion in revenue, over 50,000 ChatGPT-referred transactions<\/a>found LLM traffic converts above paid social but below every other traditional channel, and does not yet function as a broad conversion channel. Use that one.<\/div><\/div>\n\n\n\n
TAKEAWAY<\/strong>
Zero-click search has accelerated to 68% of US Google searches, but AI referrals remain around 0.14% of all web visits. The widely quoted conversion advantages for AI traffic come from single-site samples and are contradicted by the largest peer-reviewed study available.<\/div><\/div>\n\n\n\n

Answer engines in the Gulf and Egypt<\/h2>\n\n\n\n

Everything above is the global picture. Here is what changes when your market is Riyadh, Dubai or Cairo, and this section does not exist in any competing guide on this term.<\/p>\n\n\n\n

Capability<\/th>Status in Egypt and the GCC<\/th>Source<\/th><\/tr><\/thead>
AI Overviews<\/td>Live<\/strong> in Egypt, Saudi Arabia, the UAE, Qatar, Kuwait, Bahrain and Oman. Arabic is a supported language.<\/td>Google Search Help<\/a><\/td><\/tr>
AI Mode<\/td>Live<\/strong> across the same markets, with Arabic supported.<\/td>Google Search Help<\/a><\/td><\/tr>
Gemini 3 Pro inside AI Mode<\/td>English only.<\/strong> Egypt and the Gulf are on the eligible country list, but the frontier model is gated to English.<\/td>Google Search Help<\/a><\/td><\/tr>
Turning AI Overviews off<\/td>Not possible. They are a core Search feature; only the post-search Web filter removes them.<\/td>Google Search Help<\/a><\/td><\/tr>
Apple Siri in Arabic<\/td>Not supported.<\/strong> Arabic is absent from Apple’s Richer Language Understanding, All New Design, Product Knowledge and Type to Siri lists. Arabic support is input-only: dictation and voice control.<\/td>Apple<\/a><\/td><\/tr>
Arabic-language models<\/td>Falcon-H1 Arabic (UAE), ALLaM (Saudi Arabia), Jais (UAE) and Fanar (Qatar) are all live and state-backed. Fanar includes a customized Islamic RAG system and dialect-aware speech recognition.<\/td>Vendor and academic publications<\/td><\/tr><\/tbody><\/table>
Verified against official documentation on 6 August 2026.<\/figcaption><\/figure>\n\n\n\n

What this means practically<\/h3>\n\n\n\n