Become the source answer engines quote
Answer Engine Optimization makes a page directly quotable inside search answer boxes and AI assistants. We inventory the questions your buyers actually ask, reduce each answer to an extractable passage, write the markup, and measure the outcome with a date stamp. This page is the method, applied to itself.
SCOPE STRIP
Whose job is it to be quoted by an answer engine?
This work sits between the team writing the content and the team coding the page. To quote a passage, an answer engine reads both the question–answer structure of the text and the markup around it. A marketing agency does not write markup; a software agency does not build a question inventory. Separated, AEO belongs to nobody.
- Answer Engine Optimization (AEO)
- AEO is the work of making a page directly quotable inside search answer boxes and AI assistants. SEO targets ranking position; AEO targets being the source of the answer.
- Why it matters: if the buyer receives the answer on the results page or inside an assistant, a ranking position on its own no longer produces a visit. The unit of visibility stops being the page and becomes the passage.
- AEO guide Glossary definition
| Responsibility | Marketing agency | Software agency | Baki Bilisim |
|---|---|---|---|
| Who decides which questions get answered? | Partly | No | Yes |
| Who writes the FAQPage, HowTo and DefinedTerm markup? | No | Partly | Yes |
| Who measures the quoted passage and reports it with a date stamp? | No | No | Yes |
The gap is produced by the division of labour, not by a single page. The full comparison sits on why we publish measurements instead of claims.
Why doesn't ChatGPT quote my site?
One of three technical reasons applies: the page cannot be crawled, cannot be parsed, or contains no quotable passage. Answer engines do not mine meaning out of long paragraphs; they look for a short, self-contained block whose position is obvious. Without that block a page is skipped as a source even when it is indexed.
REASON 01
The page cannot be crawled
AI crawlers arrive under their own user agent names: GPTBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot, Applebot-Extended. On most corporate sites those names have never been considered at all, and some firewall or bot-protection rules block them silently. Without access, content quality is irrelevant — nothing is read.
Allowing or blocking them is a strategy decision, and the decision belongs to your company. Our job is to put both outcomes in writing and then verify on the server side that the policy you chose is the policy that is actually enforced.
REASON 02
The content cannot be parsed
If content only appears after JavaScript executes in the browser, the HTML the server returns is empty. Text buried inside tabs, accordions and carousels; paragraphs enlarged visually instead of marked as headings; a broken heading hierarchy — all produce the same result. The machine sees words but cannot tell which question they answer.
REASON 03
There is no quotable passage
This is the most common reason, and it is editorial rather than technical. The answer sits in the third paragraph, halfway through a sentence, dependent on the paragraph before it. There is no block that stands on its own for an engine to copy and attribute. On corporate sites it usually begins: “With its long-standing experience in the sector, our company…”
For exporters there is a fourth variant of the same fault: the answer exists, but only inside a PDF catalogue or a specification sheet that no answer engine will summarise on your behalf.
The three reasons do not cancel each other out; we do not guess which one applies, we check. The free digital asset audit reports the measurable part of the first two: the robots.txt rule for each crawler, a single h1 and question-form headings, and the share of home page text delivered in the raw HTML. Server-side blocks and passage structure are assessed in the AEO engagement's own review.
August 2023
Google restricted FAQ rich results to well-known government and health sites. Markup alone no longer produces a visible gain; the gain comes from the structure of the answer.
Source: Google Search Central — FAQPage structured data documentation · Accessed: 2026-07-29 · Open the documentation (opens in a new tab)
What exactly is delivered in an AEO engagement?
Six items are delivered: a priority-scored question inventory, a page-level answer-first restructuring plan, the written extractable passages, FAQPage, HowTo and DefinedTerm markup, an internal anchor architecture, and a date-stamped measurement set-up. The format and delivery week of each item is written into the contract. You receive files, not a presentation.
- Extractable passage
- An extractable passage is a 40–60 word block that answers one question without needing the rest of the page. An answer engine takes its quote from that block, not from the page as a whole.
- Why it matters: every “Answer” block on this page is written to that rule. Having applied the method rather than described it is the only claim here that you can audit yourself.
- Glossary definition Passage writing rules
| # | Deliverable | Format | When |
|---|---|---|---|
| 01 | Question inventory — compiled from search queries, the questions your sales team hears and your support records, then scored by purchase intent | CSV + priority score | Week 2 |
| 02 | Answer-first restructuring plan — which question is answered on which page, under which heading | Page-level brief | Weeks 3–4 |
| 03 | Extractable passages — answer blocks written to the length, position and self-containment rule, per language on multilingual sites | Edited page copy | Weeks 4–8 |
| 04 | Structured data — FAQPage, HowTo and DefinedTerm markup, identical to the visible content | JSON-LD + validation report | Weeks 6–9 |
| 05 | Internal anchor architecture — an anchor for every answer, terms linked to the glossary, duplicated coverage closed | Anchor map + implementation | Weeks 7–10 |
| 06 | Measurement set-up and baseline — question set, sampling method and the first record | Question set + date-stamped record | Weeks 10–12 |
What does deliverable 04 look like?
The block below is a shortened version of this page's own FAQ markup. The rule is simple: the text in the schema has to be the same text a visitor can read on the page.
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "How does AEO differ from SEO?",
"acceptedAnswer": {
"@type": "Answer",
"text": "SEO targets where a page sits in the
result list; AEO targets whether it is
selected as the source of the answer."
}
}]
}Building the entity graph behind that markup is a separate discipline: Schema.org structured data.
In what order is AEO work carried out?
In five stages: question inventory, answer architecture, passage writing, markup and measurement. The order is fixed, because markup cannot mark up an answer nobody has written yet, and work without a baseline cannot be reported afterwards. The first cycle typically runs 8–12 weeks, then narrows and repeats for new question groups.
-
01
Question inventory
Search queries, the questions your sales team hears on calls and the records in your support inbox are merged into a single list. Each question is prioritised by purchase intent and by the expertise your company can genuinely demonstrate — not by search volume.
-
02
Answer architecture
Every question is assigned to one page and to one heading on that page. A scope and anchor map is produced so the same question is never answered on two pages; duplicated coverage splits the reader and the engine alike.
-
03
Passage writing
A self-contained 40–60 word answer block is written directly beneath each heading. The definition arrives in the first sentence; figures, dates and thresholds carry their source and access date. On multilingual sites each language is written, not machine-translated.
-
04
Markup and machine readability
FAQPage, HowTo and DefinedTerm markup identical to the visible content is added as JSON-LD. The access policy for AI crawlers is settled as a written decision in robots.txt and then verified on the server side rather than assumed.
-
05
Measurement and repetition
A baseline is taken across the agreed question set, recorded with a date stamp and repeated periodically with the same questions. What gets reported is not a claim but the difference between two measurements.
These five stages sit inside the six-stage working model, contract and exit terms; no separate governance rhythm is introduced.
Which thresholds decide whether AEO work is accepted?
Six criteria: answer-block coverage, structured data error count, AI crawler access, brand citation sampling, answer box visibility and click share. The first three carry hard thresholds and are conditions of delivery. The last three are sampled, so change against a baseline is reported instead of a fixed target.
| Criterion | Threshold / acceptance condition | Measurement tool |
|---|---|---|
| Answer-block coverage | A self-contained answer block on the page for every question in the priority set | Content audit + passage checklist |
| Structured data validity | 0 errors | Schema.org Validator, Google Rich Results Test |
| AI crawler access | A written, verified decision in robots.txt for every targeted crawler | robots.txt audit + server log check |
| Brand citation sampling | Change against baseline; no guarantee is given | Periodic manual and automated checks on the agreed question set |
| Answer box visibility | Change against baseline | Search Console + results page sampling |
| Click share (CTR) | Change against baseline | Google Search Console |
We do not undertake that an AI assistant will cite your brand for a given question. Generative model output is not deterministic: the same question can produce different answers on the same day. That is why measurement is a repeated, date-stamped sample across a fixed question set rather than a single screenshot. We do not promise rankings either.
Our own site's measured results, with the tool and the date beside each figure, are published on this site's measured scores. Against the “AEO-ready” label sold elsewhere in the market, our position fits in one sentence: not a label, a measurement.
Where does AEO pay for itself first?
In sectors where the buyer arrives with a long list of technical questions. In manufacturing, retail chains, dealer networks, export industry and holding structures, dozens of specific questions precede a purchase: capacity, certification, lead time, service coverage, compliance. If your site has no answer, another source supplies one.
- 01 Manufacturing and Factories Technical buyers type capacity and tolerance questions into a search box; a manufacturer whose answers live only inside a catalogue PDF is not in that conversation.
- 02 Retail Chains Questions such as which branch stocks an item, until when it is open and how returns work are answered directly in local search and in assistants; being the source sends the visit to the store.
- 03 Franchise Networks If head office does not answer a prospective franchisee's questions about investment, support and contract terms, forums and third-party content will answer them instead.
- 04 Export Industry Overseas buyers increasingly ask an assistant to draft the supplier shortlist; if your answers are not readable by machines in the buyer's language, the shortlist is drawn up without you.
- 05 Corporate Holdings When the relationship between a holding and its subsidiaries is not marked up clearly, engines describe the group incorrectly; entity clarity is corporate reputation as machines read it.
Eight questions we are asked most about AEO
What is AEO?
AEO (Answer Engine Optimization) is the work of making a page directly quotable inside search answer boxes and AI assistants. It targets being the source of the answer rather than a ranking position. In practice it involves three jobs: inventorying the questions buyers ask, reducing each answer to a short self-contained passage, and exposing that structure to machines through FAQPage, HowTo and DefinedTerm markup.
How does AEO differ from SEO?
SEO targets where a page sits in the result list; AEO targets whether that page is selected as the source of the answer. The unit of SEO is the page and the query; the unit of AEO is the passage and the question. They are not rivals: SEO builds the technical base of crawlability, information architecture and page speed, and AEO adds the answer structure and the markup on top of it.
When do ChatGPT and Perplexity quote a page?
When three conditions hold at once: the page is reachable, meaning AI crawlers are not blocked in robots.txt or at the server; the content is readable in the HTML that the server returns rather than only after JavaScript runs; and the answer to the question exists as a short, self-contained passage. If any one of the three is missing, the page is not selected as a source even when it is indexed.
How is an extractable passage written?
The question is asked in a heading, the answer is given in the first sentence beneath it, and the passage is kept to 40–60 words. It must make sense without the rest of the page: the subject is stated explicitly and phrases that depend on earlier context are avoided. Figures, dates and thresholds carry their source and access date. On multilingual sites each language gets its own passage, because a machine-translated answer reads as a translation and is quoted less.
How are AEO results measured?
With four measurements: answer-block coverage across the target question set, the error count in a structured data validator, brand citation sampling across an agreed question set, and impressions plus click share in Search Console. Sampling is recorded with a date stamp and repeated with the same questions, per language on multilingual sites. No citation guarantee is given; what is reported is the difference between two measurements.
Is adding FAQ schema enough?
No. In August 2023 Google restricted FAQ rich results to well-known government and health sites, so the markup on its own no longer produces a visible gain. Schema makes visible content easier for machines to read, but if the content was not written answer-first there is no passage to quote in the first place. The order is content first, markup second.
Can AEO work damage my existing SEO?
No; done properly it pulls in the same direction as SEO. The risk sits in changing URL and heading structures without reason. That is why the work preserves existing URLs, sets up a 301 redirect for every address that genuinely has to change, and takes a page-level baseline before any content is edited, so that any movement can be attributed.
How long before results appear?
The first cycle typically runs 8–12 weeks: question inventory, answer architecture, passage writing, markup and baseline measurement. How quickly structural changes surface in search depends on how often the pages are recrawled. Citation inside AI assistants depends on each model's data refresh rhythm, so we give no date for it — which is precisely why the measurement is repeated rather than taken once.
Questions about process, pricing, contracts and international invoicing are answered on the general FAQ page.
See the measured state of your website within five working days.
The audit is free and creates no obligation to work with us. The report itemises findings on AEO answerability, AI crawler access, lab-measured Core Web Vitals, structured data validity and accessibility (automated scan).
We work with corporate-scale, multi-location or multilingual organisations. One-off small jobs fall outside our scope; in that case we point you to smaller studios.