01GEO
GEO: generative engine optimisation, minus the mythology
GEO is optimisation for AI-generated summaries inside search engines, principally Google AI Overviews and AI Mode. Searlixa runs GEO as a measurement-led programme built on core search strength, because that is what the evidence supports.
Quick answer
GEO is optimisation for AI summaries generated inside a search engine, principally Google AI Overviews and AI Mode.
Google states this is still SEO, and names llms.txt, chunking and AI-specific schema as unnecessary. We agree. GEO exists as a distinct service because the outcome is now measurable and no longer follows automatically from ranking.
Key takeaways
- Most of the GEO market sells tactics Google has publicly dismissed. Treat a proposal containing them as a signal about the firm making it.
- Ranking is the biggest lever here, more than on any other AI surface.
- Trigger rate is its own metric. A query that stops producing an Overview has not removed you from anything.
- Answer in the first 100 words. 22.1% citation rate against 8.7% otherwise.
- Guarantees are impossible here. Generative answers are non-deterministic.
- GEO and AEO are one programme for most clients, not two.
Definition
Generative engine optimisation (GEO) is the practice of improving how often and how accurately a brand is cited inside AI summaries generated by search engines. Google's position is that this is still SEO. Our position is that the optimisation is, but the measurement is not.
Start with what Google actually says
In May 2026 Google published guidance on optimising for its generative AI features. It states that optimising for AI search is optimising for the search experience, and therefore still SEO. It explicitly names things you do not need: llms.txt files, machine-readable AI text files, content chunking, AI-specific structured data, or content rewritten "for AI".
A large part of the GEO market is selling exactly those things. We are not going to, and you should treat any proposal that includes them as a signal about the firm making it.
So why does GEO exist as a service at all?
Because the outcome is now measurable and it was not before, and because the outcome no longer follows automatically from ranking. Ahrefs measured the share of AI Overview citations coming from top-10 organic results falling from roughly 76% in July 2025 to 38% by March 2026. You can hold position three and be absent from the summary above it.
What we actually do
- Baseline the summaries. Sample a frozen buyer prompt set against AI Overviews and AI Mode, record whether a summary triggers, whether you are cited, and who is cited instead. Trigger rate is tracked as its own metric, because a query that stops producing an Overview changes the funnel without anything being wrong with your page.
- Fix the ranking gateway. Traditional top-three positions appeared in AI citations 98.9% of the time in the 2026 study. Most GEO budget is best spent here, and that means ordinary search work done properly.
- Broaden query coverage per page. Pages serving eight or more related queries reached the AI top ten 43.3% of the time against 15.5% for single-query pages. Deepening beats publishing.
- Front-load the answer. Pages introducing their topic within the first 100 words were cited at 22.1% against 8.7% otherwise: the strongest structural signal measured.
- Raise factual specificity. Named entities, dated figures, stated sources. Specific pages outperformed vague ones consistently.
- Clear the technical path. Rendering, indexation and a deliberate crawler policy, handled through crawl and render work.
What we will not do
| Tactic | Our position | Basis |
|---|---|---|
| llms.txt file | Will explain it, will not bill for it | Google states it receives no special treatment; adoption is under 10% of the top 1,000 sites |
| Content chunking | Not recommended | Google states there is no ideal length and no need to fragment pages |
| AI-specific schema | Not sold as a lever | No schema type showed a consistent independent citation advantage in the 2026 study |
| Guaranteed AI Overview placement | Impossible to offer honestly | Generative responses are non-deterministic; the same prompt varies between runs |
| Bought or seeded mentions | Refused | Inauthentic mentions are named in Google's spam policies, which now cover generative responses |
| Rewriting pages "for AI" | Not a thing we do | Google names it directly as unnecessary. Write clearly for readers and the structural signals follow |
GEO or AEO?
GEO targets summaries generated inside a search engine. Answer engine work targets assistants that answer instead of returning a results page. The work overlaps heavily, and most clients need one programme rather than two.
If you are unsure which surface you are losing on, that is what the two-week baseline answers before you commit budget to either. You can also run a rough version yourself — the three checks for AI Overview absence take an afternoon.
02Common questions
What buyers ask about GEO
Is GEO a real discipline or just SEO rebranded?
Google says it is still SEO, and for the optimisation work that is broadly right. What is genuinely new is the measurement: nothing in Search Console tells you whether an AI summary cites you, so GEO exists as a distinct service for that reason rather than for its tactics.
Do I need an llms.txt file for GEO?
No. Google states the file receives no special treatment and adoption sits under 10% of the top 1,000 websites. We will explain it and help you decide, but we will not bill for implementing it as a visibility service.
How is GEO different from AEO?
GEO targets AI summaries generated inside a search engine, principally Google AI Overviews and AI Mode. AEO targets assistants that answer instead of returning results. The work overlaps heavily and most clients need one programme, not two.
Can you guarantee an AI Overview placement?
No. Generative responses are non-deterministic, so the same query can produce different sources on consecutive runs. Any guarantee would be a promise about a system nobody controls, and it should make you suspicious of whoever offers it.
What is the single biggest lever?
Ranking, still. Traditional top-three pages appeared in AI citations 98.9% of the time in 2026 research. It is no longer sufficient, since more than half of citations come from outside the top ten, but it remains the strongest predictor available.
How long until we see movement?
Structural fixes such as answering earlier on a page can show within weeks on this surface, because Google re-crawls frequently. Coverage and authority gaps take months, and on a new domain in a contested category, twelve to eighteen.
→What you are buying
GEO services: what is included
GEO agency work is presence inside generative experiences, measured per platform because the platforms disagree with each other.
- Presence measured inside generative experiences, per platform and per market
- Content shaped so it can be used as a source rather than summarised around
- Entity accuracy work, because being named wrongly is worse than being absent
- Reporting that separates trigger rate from citation presence
Who this is for
Brands in categories where AI summaries already appear on most queries.
Who it is not for
Anyone expecting a ranking position inside a generative answer. There is no position — citation is binary.
From $900 a month. Scoped together with AEO, billed once. Buying them separately means paying twice for one body of work. SEO and AI visibility pricing is published in full.
The audit checks generative presence on your priority pages across every platform we support. The $99 SEO and AI visibility audit comes back in three working days, and the fee comes off your first invoice if you continue.
03Related
Continue
The umbrella programme · How AI Overviews cite · SEO · Measure it yourself · Pricing
Sources
Every figure on this page comes from one of the following. Where a claim is ours rather than a source’s, it carries an evidence label instead.
- Google Search Central — Optimizing your website for generative AI features (published 15 May 2026, updated 10 July 2026). Source for the llms.txt, chunking and AI-schema positions, and for the statement that generative features run on core ranking systems. Primary
- Ahrefs — AI Overview citation analysis, March 2026. Source for the fall in top-10 organic share of citations from roughly 76% to 38%. Independent study
- Rankability — State of AI Search, 1,645 AI citation observations across 31 topics, 2026. Source for the 24.1% cross-platform overlap, the 73.5% ChatGPT exclusive citation rate, the 22.1% vs 8.7% first-100-words finding and the 43.3% vs 15.5% query-breadth finding. Independent study
- Google Search Central — spam policies, now explicitly in scope for generative AI responses in Search. Primary
- Searlixa sampling across client engagements. Any figure drawn from this carries an Observed label with its sample size and date. Our own
Ask an assistant
Hand this page to an assistant and ask it to summarise or challenge the argument. If it gets something wrong about us, that is itself a useful finding.