AI Search (AEO)

What AI search and Answer Engine Optimisation actually mean

AI search refers to the way large language models and AI-powered tools now answer questions directly, often citing or summarising sources rather than sending people to a list of blue links. You see it in Google's AI Overviews, in ChatGPT (especially when it browses the web), in Google Gemini, in Microsoft Copilot, and in Perplexity, which was built around the idea of answering with cited sources.

Two terms get used for the practice of earning visibility in these answers:

• AEO (Answer Engine Optimisation) - making your content easy for an answer engine to quote, summarise and attribute to you.
• GEO (Generative Engine Optimisation) - the same idea framed around generative AI results specifically.

In practice the two overlap heavily, and we treat them as one discipline. The goal is straightforward: when someone asks an AI tool a question your business can answer, you want your content to be the thing it draws on, and ideally the source it names.

This is a genuinely new and fast-moving area. The platforms change their behaviour often, citations are not always stable, and there is no published ranking formula. We work from observable patterns and tested principles rather than promises, and we will tell you plainly where the evidence is thin.

How AI search relates to traditional SEO

The most important thing to understand is that AEO is not a separate channel you bolt on. It builds directly on the foundations of good SEO. AI systems are trained on, and frequently retrieve from, the same web that search engines crawl and index. Google's AI Overviews draw on its existing index. Tools like ChatGPT and Perplexity use web search behind the scenes to ground their answers, which means crawlable, indexable, well-structured pages are still the entry point.

So the work that earns you organic rankings - technically sound pages, clear information architecture, genuinely useful content, and a credible link profile - is largely the same work that makes you citable by AI. If your SEO fundamentals are weak, no amount of AI-specific tactics will compensate. If they are strong, AEO becomes an extension of what you already do, with some added emphasis on how content is written and structured for extraction.

The practical implication: invest in AEO as a layer on top of a healthy SEO programme, not as a replacement for it.

What actually influences whether an AI tool cites you

There is no confirmed list of ranking signals for AI answers, but the patterns we see and the principles the platforms describe point consistently in a few directions.

• Clear, extractable structure. Content that directly answers a question, near the top, in plain language, is easier for a model to lift and attribute. Descriptive headings, short self-contained paragraphs, and sensible use of lists and tables all help.
• Genuine expertise and E-E-A-T. Experience, expertise, authoritativeness and trustworthiness matter more in AI answers, not less. Content that demonstrates real first-hand knowledge, names its authors, and is accurate tends to be favoured over thin, generic pages.
• Well-defined entities. AI systems reason about people, organisations, products and places as entities, not just keywords. Being consistently and accurately described across your site and the wider web helps a model understand who you are and what you are an authority on.
• Structured data. Schema markup gives machines an unambiguous description of your content. It does not guarantee a citation, but it removes ambiguity and supports the broader job of being understood.
• Authority and mentions. Being referenced, quoted and linked by other credible sources signals that you are worth citing. AI tools appear to lean on sources that the wider web already treats as authoritative.
• Freshness and accuracy. For questions where currency matters, up-to-date and factually correct content is more likely to be surfaced and trusted.

None of these are tricks. They are the same things that make content valuable to a human reader, expressed in a way machines can parse.

How AI visibility is measured

Measurement in this space is less mature than traditional rank tracking, and honesty about that matters. AI answers can vary between users, sessions and even identical prompts, so you cannot expect the clean, repeatable position data you get from organic search.

What can be tracked, with reasonable rigour:

• Citation and mention monitoring - checking, across a defined set of representative prompts, whether your brand or pages appear in answers from tools like ChatGPT, Gemini, Perplexity and AI Overviews.
• Share of voice against competitors - how often you are cited relative to others for the questions that matter to your market.
• Referral traffic from AI sources - identifying visits from AI tools in your analytics, which is improving but still imperfect to attribute.
• Coverage of priority questions - whether the questions your customers actually ask are answered well on pages you control.

We set this up as a repeatable check rather than a one-off snapshot, so you can see direction of travel over time. We are also clear that these metrics are indicative, not exact, and we report them that way.

Who should invest now, and who can wait

AEO is worth prioritising sooner if your audience already uses AI tools to research before they buy, if your category involves considered or complex decisions where people ask detailed questions, or if you have strong subject-matter expertise that is currently under-represented online. In those cases, being citable now builds an advantage while the field is still forming.

If your SEO foundations are not yet in place, the better sequence is usually to fix those first, because they do most of the heavy lifting for AI visibility anyway. For many businesses the right move is a steady, integrated approach: keep doing sound SEO, add the structural and authority work that supports citability, and measure as you go - rather than chasing every new platform feature as it appears.