What is generative engine optimization (GEO)?

Generative engine optimization (GEO) is the work of making a business easy for AI assistants such as ChatGPT, Gemini, Perplexity and Copilot to find, understand and cite when they answer a question. It builds on SEO instead of replacing it: AI answers are assembled from pages that search systems can already retrieve, so being retrievable comes first and the wording of the page second.

Updated 23 October 2026 · By AYKA TECHNOLOGIES

GEO vs SEO: what changes and what stays the same

SEO aims for a position in a list of results and a click. GEO aims for a mention or a citation inside an answer. Both start from the same crawl of your pages, which is why a page that does poorly in search rarely gets cited by an AI assistant.

Why does it matter now?

AI assistants increasingly decide which businesses a customer even considers, and buying is moving into the same conversations. Morgan Stanley estimates that AI shopping agents could drive $190 billion to $385 billion of US e-commerce spending by 2030 (Morgan Stanley Research). A business that assistants cannot find or describe correctly is left out of those answers.

How do AI assistants choose which sources to cite?

For one question, an assistant usually runs several searches, reads the passages that answer each part, and writes an answer with links to the pages it used. Google describes this for AI Overviews and AI Mode as a “query fan-out” technique, issuing multiple related searches across subtopics and data sources, according to its Search Central documentation.

1. Retrieve. The assistant searches its index for each sub-question. A page that is blocked, not indexed or ranked far down is not in the running.
2. Select. It keeps passages that answer a sub-question directly, with specific facts it can repeat.
3. Compose. It writes the answer and links the sources it relied on.

The practical consequence: one strong page per topic that answers the main question and its follow-ups beats many thin pages that each repeat one phrasing.

What does the research say works?

The study that named the field tested content changes across thousands of queries. Its authors (Aggarwal et al., KDD 2024) report:

“GEO can boost visibility by up to 40% in generative engine responses.”

The changes that helped most were adding citations to credible sources, quotations and specific statistics. Stuffing the page with the query’s keywords did not help.

A later benchmark adds an important limit. C-SEO Bench (Puerto et al., NeurIPS 2025) found that “most current C-SEO methods are not only largely ineffective but also frequently have a negative impact on document ranking”, that classic SEO which improves a page’s ranking works better, and that the gains shrink as more sites use the same tricks. In short: wording tricks are not a strategy. Accurate, well-sourced pages that rank are.

• Do: answer first, cite your sources, give specific numbers with dates, name the people or organisations you quote, keep one page per topic.
• Avoid: hidden text, sentences addressed to AI models, invented statistics, and pages that only repeat a keyword in new orders. Hidden text and scaled, low-value pages are named in Google’s spam policies.

Do llms.txt files or special markup help?

Not for Google. Its documentation is direct:

“There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.”

The same page says you do not need new machine-readable files, AI text files or special schema.org markup. Google has also said that llms.txt files will neither harm nor help visibility in Google Search, because Search ignores them (Search Engine Land). Other services may read them, so we publish one, but we never sell it as a ranking factor.

Structured data still has a job: it states your name, address, logo and profiles in one consistent, machine-readable form. Google’s rule is that it must match the visible text of the page.

GEO in Arabic: what is different?

Customers in the UAE ask the same question in Arabic, in English, and in a mix of both. A page that answers only in English cannot be quoted for the Arabic question. What we do for Arabic pages:

• Write the Arabic page natively, in clear Modern Standard Arabic, instead of a word-for-word translation.
• Use the words customers type, including English brand and service names that are commonly written in Latin letters.
• Name the area in both languages (Al Barsha / البرشاء) and keep one Arabic spelling of your business name across your site, Google Business Profile and directories.
• Give the Arabic and English pages the same facts, prices and dates, so an assistant never finds two versions of the truth.

How long does GEO take to work?

It depends on how fast your pages are crawled and indexed. Changes to pages that are already indexed can show up in answers within weeks; a new website usually needs months to earn rankings first. AI answers also change from one run to the next, so a single check proves little. We measure over a fixed set of buyer questions every month and compare them with control questions we never optimise, so the report shows our effect, not noise.

What does a GEO project for a Dubai business include?

1. Audit: your customers’ questions in Arabic and English, put to ChatGPT, Gemini, Perplexity and Copilot; who they name and which sources they trust.
2. Foundation: crawler access, indexing, Google Business Profile, and one consistent company record (name, address, phone, hours, profiles) on every site that describes you.
3. Pages: answer-first pages for the topics your buyers ask about, with sources, dates and the follow-up questions covered.
4. Measurement: a monthly report of test questions against control questions.

No one can guarantee what an AI assistant will say. We report what changed and why.