AI SEO: Google ranking and AI answers are measured together
AI SEO makes a page visible both in Google's organic ranking and in the answers given by ChatGPT, Perplexity and Google AI Overviews. Generative engine optimisation, known as GEO, aims to get a page cited inside an AI answer. Answer engine optimisation, known as AEO, gives a page a short answer that can be quoted on its own. ALTAI Digital measures each engine separately and makes no claim for one it has not measured.
A business can sit near the top of Google and still go unmentioned when someone asks ChatGPT the same question. The two surfaces run on different engines. They use different ranking rules and need separate measurement. AI SEO treats both together. Google and Bing's organic ranking and the AI answers from ChatGPT, Perplexity and Google AI Overviews are tracked on their own. GEO, generative engine optimisation, aims to get a page cited as a source inside an AI answer. AEO, answer engine optimisation, gives a page a short answer an AI engine can quote directly. The two work side by side; neither replaces the other.
The first signal to check is the index rate. A page that is not in an engine's index cannot rank in that engine. On the Google side, the URL inspection tool in Search Console shows this. Ranking change is never read from a single day. It is compared across a twenty eight day window against the twenty eight days before it. A single day of movement proves nothing on its own, and this window keeps that noise out of the picture.
Our own measurement (the altai.digital market, 23 September 2026) looked at advice style queries. Questions phrased as what should I pick or what is best returned AI answers. Those answers cited listicles and directories far more than vendor sites. The same pattern showed up on three separate surfaces, in an AI chat answer, in Bing and in Perplexity. A plain commercial query in the same market still returned a standard organic list. That split points to two tracks of work: improving the page itself, and getting listed in the sources an AI engine already cites.
The same measurement found that ChatGPT answers some local business questions with a map style card. The card shows a rating and a review count. That makes a Google Business Profile part of AI SEO, not a separate task limited to map apps. The technical layer matters just as much. A short answer near the top of the page, question style headings and an FAQ section with FAQPage schema all sit inside this layer. So do a robots.txt file open to AI crawlers, an llms.txt file and an IndexNow ping to Bing. ALTAI Digital works from Rotterdam and Istanbul.
What goes wrong today
- 01
Ranks on Google, absent from AI answers
A page sits near the top of Google's organic results. The same question put to ChatGPT or Perplexity never mentions it. The two engines run on different ranking rules, so one result does not carry over to the other. AI SEO tracks both separately and improves each on its own terms.
- 02
Advice queries favour listicles over vendor sites
In our own measurement, a question phrased as what should I pick returned an AI answer that mostly cited a listicle or a directory. In that measurement, the vendors' own pages were mostly not the cited sources. The fix runs two tracks: improving the page, and getting listed where the AI engine already looks. A plain commercial query returned an organic list.
- 03
A page outside the index is invisible everywhere
A page that is not in an engine's index cannot rank in that engine. The URL inspection tool in Search Console shows the real state directly.
- 04
No short answer for an AI engine to quote
Without a short answer near the top of the page, an AI engine has to read the whole page. It may not find a short passage to quote. AEO closes this gap by placing a short answer paragraph at the top. Question style headings and an FAQ section serve the same goal.
- 05
An AI crawler cannot reach the page
A robots.txt file that blocks AI crawlers keeps a page out of that engine's answer. This holds no matter how well the page is written. An llms.txt file and an IndexNow ping may be missing too. This gap goes unnoticed while content work keeps running.
- 06
Content ships without passing a quality check
A model can speed up a draft. The topic, the sources and the voice still have to come from the client. A quality gate checks four things: the source behind every figure and quote, brand voice, originality and technical structure. A person always makes the call to publish.
Scope of the build
Measurement split by engine
Organic ranking on Google, Bing and Yandex and AI answers from ChatGPT, Perplexity and Google AI Overviews are tracked separately. No result is claimed for an engine that has not been measured. Each engine keeps its own measurement window. This split shows which part of the work is paying off.
Index status checked first
Every page's index status is checked through the URL inspection tool in Search Console. A page that is not in an engine's index cannot rank in that engine. Ranking is compared across a twenty eight day window against the previous one. An index problem is fixed before anything else.
Generative engine optimisation (GEO)
GEO aims to get a page cited as a source inside the answers from ChatGPT, Perplexity and Google AI Overviews. Advice queries favour listicles and directories. So GEO covers both strengthening the page itself and getting listed where an AI engine already looks. A plain commercial query may not need this extra track.
Answer engine optimisation (AEO)
AEO builds a short, quotable answer near the top of the page. It adds question style headings, an FAQ section and FAQPage schema. This structure makes it easier for an AI engine to quote the page directly. The short answer paragraph sits at the very top of the page.
Technical access layer
A robots.txt file open to AI crawlers, an llms.txt file and an IndexNow ping to Bing sit inside this layer. So do a sitemap and a lastmod date that updates with the content. These settings speed up discovery and notification. This layer is checked before content work starts.
Google Business Profile and the map card
ChatGPT answers some local business questions with a map style card showing a rating and a review count. That makes a Google Business Profile part of AI SEO, not a task limited to map apps. Keeping that profile accurate keeps the card itself accurate. This gets tracked separately for local queries.
Content quality gate and inventory review
A model speeds up the draft; the topic, the sources and the voice come from the client. The quality gate checks the source behind every figure, brand voice, originality and technical structure. Existing content is sorted into three groups: update, merge or keep as is. A person makes the call to publish in every case.
How the build runs
- 01
Audit and index check
Index status is checked for existing pages, and the content inventory is pulled together. Each page is sorted into update, merge or keep. The technical layer is reviewed.
- 02
Technical build
A robots.txt file open to AI crawlers, an llms.txt file, an IndexNow ping and FAQPage schema are put in place. A short answer and question style headings are added to pages.
- 03
Draft and quality gate
Topic, sources and voice come from the client; a draft is prepared with the model's help. The quality gate checks sources, brand voice and structure. A person signs off before publishing.
- 04
Measurement by engine
Organic ranking and AI answers are tracked separately. Results are compared across twenty eight day windows. No claim is made for an engine that was not measured.
Products used in this build
Terms used on this page
- GEO
- GEO, short for generative engine optimisation, aims to get a page cited as a source inside the answers given by AI engines such as ChatGPT and Perplexity.
- AEO
- AEO, short for answer engine optimisation, gives a page a short, self-contained answer that an AI engine can quote directly.
- Index rate
- The index rate is the share of a site's pages that search engines and AI crawlers have reached and stored. It is checked through the URL inspection tool in Search Console.
- llms.txt
- An llms.txt file sits at the root of a site. It tells AI models which content and links to treat as the priority source.
- FAQPage schema
- FAQPage schema is a structured data type that marks up a page's questions and answers. It lets search and AI engines read them directly.
AI SEOFrequently asked questions
AI SEO makes a page visible in Google and Bing's organic ranking. It also covers the AI answers from ChatGPT, Perplexity and Google AI Overviews. The two surfaces run on different engines, so they are measured separately. ALTAI Digital makes no claim for an engine it has not measured.
GEO, generative engine optimisation, aims to get a page cited as a source inside an AI answer. It is one part of AI SEO, not the whole of it. A GEO agency works on getting a business cited in AI answers. ALTAI Digital runs GEO alongside organic ranking on Google and Bing.
Inside AI SEO, AEO, answer engine optimisation, gives a page a short answer an AI engine can quote directly. A short answer paragraph, question style headings, an FAQ section and FAQPage schema are the tools behind it. Without AEO, even accurate content can be harder for an AI engine to quote.
No. Our own measurement found that advice style queries returned AI answers citing listicles and directories far more than vendor sites. That pattern showed up on three separate surfaces. A plain commercial query still returned a standard organic list. That is why the two surfaces are worked on separately.
A page that is not in an engine's index cannot rank in that engine. Index status is checked through the URL inspection tool in Search Console. Skipping this step leaves later content work with nothing to build on. Ranking change only means something once index status is confirmed.
ChatGPT answers some local business questions with a map style card showing a rating and a review count. That is why a Google Business Profile is part of this work. Keeping that profile accurate keeps the card itself accurate.
A model speeds up the draft, but the topic, the sources and the voice come from the client. The quality gate checks the source behind every figure, brand voice, originality and technical structure. A draft that fails the gate stays unpublished, and a person makes the call to go live.
Ranking change is read across a twenty eight day window, compared with the twenty eight days before it. A single day of movement proves nothing on its own. An early date is not a fair thing to promise. Each engine is tracked in its own window, and reporting follows that window.
The process first, the proposal after
In a short discovery call we map the process together and tell you plainly whether it is worth automating. If it is not, we say that too.
30 minutes · no preparation needed
