GEO · AEO · Ireland

AI search optimisation

Most searches now end without a click. AI search optimisation, the work behind GEO and AEO, makes sure the answer people do see carries your name and links back to your site.

The shift

Search stopped sending the clicks it used to.

SparkToro's clickstream study with Similarweb found 68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% two years earlier. Google still answers the question. The click just never reaches your site.

Those last two numbers are the whole argument. Being cited inside the answer more than doubles what you get from a search that shows an AI Overview. Being left out costs you the visit twice over.

The letters

Generative engine optimisation, answer engine optimisation and SEO, without the jargon.

All three run on the same foundations: a fast, crawlable site, clear writing, structured data and a business the wider web actually mentions. We run them as one plan rather than selling you three.

What gets cited

Ranking first is no longer the entry ticket.

Ahrefs looked at 863,000 keywords and four million URLs. In early 2026 only 38% of the pages cited in AI Overviews were ranking in the organic top 10, down from 76% in mid 2025. Roughly a third of citations went to pages ranking 11 to 100, and another third to pages barely ranking at all.

Models pick sources on clarity, structure and freshness as much as position. Seer Interactive found around half the content cited in AI answers was less than 13 weeks old. Research from Princeton and IIT Delhi across 10,000 queries found that adding cited sources, direct quotes and hard statistics lifted citation rates by up to 40%.

A page that answers the question in its first 40 words, backs it with a number and names its source gets quoted. A page that warms up for four paragraphs gets skipped.How the work actually breaks down

The work

What AI search optimisation involves.

01

Entity groundwork

Your business, your people and your services described the same way everywhere, marked up with LocalBusiness, Service, Person and FAQ schema so machines know what you are, not just what you sell.

02

Answer first content

Pages that lead with a direct answer in plain English, then back it up. Short, quotable blocks. Real numbers. Tables and lists a model can lift cleanly.

03

Freshness and depth

Guides kept current with dates that mean something, because recent content wins a large share of citations. Thin pages merged rather than left to rot.

04

Mentions and citations

Coverage on Irish news sites, trade press, directories and community pages. Models lean on what other sites say about you, so unlinked mentions still count.

05

Crawler access

Checks that AI crawlers can reach your pages, that nothing important hides behind JavaScript, and that your robots rules say what you meant them to say.

06

Tracking

A fixed set of prompts run through the main assistants each month, AI Overview citations logged, and AI referral traffic split out in GA4.

Straight answer

You probably do not need an llms.txt file.

Plenty of agencies sell llms.txt as the AI search fix. Google's own AI optimisation guidance, published in May 2026, says llms.txt is not needed for AI Overviews, AI Mode or any other generative feature in Search. Anthropic and OpenAI do use the format for agents and developer docs, so it earns its place on a documentation site and does very little on a shop with 10,000 products.

We will build one where it helps. We will not bill you for one and call it a strategy.

Fit

Is this worth doing for your business?

Worth doing when your buyers research before they ring. Professional services firms whose clients ask an assistant what to look for. Irish online shops competing with UK marketplaces that already own the comparison answers. Businesses with several locations, where the assistant has to pick one to name.

Not worth doing yet on a brand new site that nothing on the web mentions. Models lean on what other sources say about you, so with no coverage, no reviews and no profile anywhere, the groundwork is ordinary SEO and digital PR first. We will tell you that rather than sell you the audit.

What we can and cannot promise

Nobody controls what a model says, and any agency promising you a citation is guessing. What we can do is measure it. We run a fixed set of prompts every month, log which sources get cited for them, and show you the prompts where a competitor is named instead of you. You see the same list we work from.

On the technical side we follow Google's own guidance on AI features, which is clear that the foundations are ordinary good SEO rather than a new file format or a secret tag. Plenty of the AI search packages being sold in Ireland contradict that, which is a useful filter when you are comparing quotes.

Who does the work

Conor O'Brien, 20 years in search.

Conor O'Brien founded Webovation and does the work himself. Twenty years in SEO means he watched search change from ten blue links to featured snippets to answers written on the fly, and the underlying job has not moved much. Understand the question, answer it better than anyone else, and make the answer easy to find and easy to trust.

Read more about Conor and how we work, or see what the wider service covers on the web design and SEO page.

Free AI visibility audit

Find out what the assistants say about you.

We run your brand and your top ten questions through Google AI Overviews, ChatGPT, Perplexity and Gemini, then send you what came back. You get the prompts where you are named, the ones where a competitor is named instead, and the three changes most likely to move that.

  • Where you are cited today, with screenshots
  • Which competitor owns the answers you want
  • The schema and content gaps behind the difference

Questions

AI search questions we get asked.

What is AI search optimisation?

AI search optimisation is the work of getting your business named and cited in AI answers, such as Google AI Overviews, ChatGPT, Perplexity and Gemini. It covers the same ground as SEO, with more weight on clear answers, structured data, entities and being mentioned on sites the models read.

What is the difference between GEO, AEO and SEO?

SEO aims at ranking a page in a list of links. AEO, answer engine optimisation, aims at being the answer shown on the page. GEO, generative engine optimisation, aims at being the source a generated answer cites. The underlying work overlaps heavily, so treat them as one plan rather than three budgets.

Does ranking first still get me cited by AI?

Ranking helps but no longer decides it. An Ahrefs study of 863,000 keywords found only 38% of pages cited in AI Overviews ranked in the organic top 10 in early 2026, down from 76% in mid 2025. Pages well outside the top 10 now earn citations on clarity and structure.

Do I need an llms.txt file?

Probably not for Google. Google's May 2026 AI optimisation guidance says llms.txt is not needed for AI Overviews or AI Mode. Anthropic and OpenAI do use the format for agents, so a small file makes sense for documentation and technical sites, and little sense for a large shop.

How do you measure AI search visibility?

We track how often your pages are cited in AI Overviews, run a fixed set of prompts through ChatGPT, Perplexity and Gemini each month to see whether you are named, and watch AI referral traffic in GA4 alongside branded search in Search Console.

Where these numbers come from

The click and citation figures are drawn from published 2026 research by SparkToro with Similarweb, Seer Interactive and Ahrefs, checked against what we see in our own clients' Search Console and GA4 accounts. Guidance on AI features and llms.txt follows Google's own documentation. We update the page whenever the research moves.