My Usual Process for AEO
It is up to you how far you want to take this. If you already know you are invisible in AI search and just need someone to implement the fixes, I can do that. If you want someone to establish where you currently stand, why, and what the plan is, that works too.
In most cases I start with a visibility audit. I run a set of real buying-intent prompts through ChatGPT, Gemini, Perplexity and Google AI Overviews, record who gets cited, how often, and which sources those answers pull from. That gives us a baseline and, more usefully, a list of the exact pages and publications the models already trust in your category.
From there the work splits into two halves. The technical half makes your site readable and quotable by AI crawlers. The content half gives those crawlers something worth citing. Both are documented in a strategy so you always know what is being built and why, and so progress can be measured against the baseline rather than guessed at.
AI Visibility Audit
The starting point for every engagement. I run your category's real buying prompts through the major assistants and record where you appear, where you do not, and who is taking the citation instead. You get a scored baseline covering ChatGPT, Gemini, Perplexity and Google AI Overviews, plus the source list each model is drawing on. Most businesses find they are absent from answers they assumed they owned.
Structured Data & Schema
Language models rely heavily on explicit, machine-readable facts. I implement and validate the schema types that matter for answer engines: Organization, Person, Service, Product, FAQPage, HowTo and Article, with the entity relationships properly connected rather than bolted on. This is the single highest-leverage technical change on most sites and it is usually the one that has been skipped.
llms.txt & Crawler Access
AI crawlers are not the same as Googlebot and are frequently blocked by default, either in robots.txt or at the CDN. I check whether GPTBot, ClaudeBot, PerplexityBot, Google-Extended and the rest can actually reach your content, fix the blocks you did not know you had, and implement an llms.txt file that gives models a clean, prioritised map of your most important pages.
Answer-Led Content
Content that ranks is not automatically content that gets quoted. Answer engines favour direct claims, specific figures, clear definitions and self-contained passages that can be lifted without losing meaning. I restructure existing pages and brief new ones so each section answers one question completely, which is what makes a passage citable in the first place.
Entity & Brand Building
Models cite sources they recognise. That recognition is built from consistent mentions across the places they are trained on and the places they retrieve from live: your own site, industry publications, directories, review platforms, Wikipedia-adjacent sources and communities. I map where your competitors are being mentioned that you are not, then work through that gap methodically. This is the slowest part of AEO and the hardest for a competitor to undo.
Tracking & Reporting
Answer engines do not hand you a rank tracker, so visibility has to be measured deliberately. I re-run the same prompt set on a fixed schedule and report on citation frequency, share of voice against named competitors, sentiment of the mention and referral traffic from AI sources isolated in analytics. You see movement against the original baseline, not a vanity score.
What is Answer Engine Optimisation?
Answer engine optimisation, or AEO, is the practice of getting your business cited inside AI-generated answers rather than inside a list of links. When someone asks an assistant for a recommendation, the model composes a single response and names a handful of sources. AEO is the work that makes yours one of them.
You will also see it called generative engine optimisation, GEO, LLM SEO or AI search optimization. The labels differ, the work overlaps heavily, and the underlying shift is the same: buyers are increasingly getting their answer without ever reaching a results page.
How is AEO different from SEO?
SEO earns you a position. AEO earns you a mention. They share a foundation, since a page an AI cannot crawl or parse is no use to either, but the objectives diverge from there. SEO rewards pages that satisfy a query well enough to be clicked. AEO rewards passages that answer a question so cleanly they can be extracted and attributed.
In practice AEO leans much harder on structured data, factual specificity, consistent off-site mentions and content architecture, and much less on the click-through tactics that drive traditional rankings. I would not treat it as a replacement for SEO. Sites that are technically sound and already trusted have a significant head start in AI answers, so the two run alongside each other. If you are weighing up providers, I have written on how to choose an AI SEO agency.
Frequently Asked Questions
How long does AEO take to show results?
Technical changes such as schema and crawler access can shift things within weeks, because models retrieve live content as well as relying on training data. Entity and citation work takes longer, typically three to six months, since it depends on third-party sources publishing and those sources being picked up. The audit gives you a baseline on day one so you can see which is moving.
Do I need AEO if my SEO is already working?
Increasingly, yes. Strong SEO gives you a real advantage in AI answers, since models lean on sources that already rank and are already trusted. But ranking first no longer guarantees you are the one named in the answer, and a competitor with better structured data and more third-party mentions can be cited above you. AEO protects the visibility your SEO has already earned.
Which AI platforms do you optimise for?
ChatGPT, Google AI Overviews and AI Mode, Google Gemini, Perplexity, Claude and Microsoft Copilot. The underlying work is largely shared across all of them, though each weights sources slightly differently, which the audit accounts for.
Can you guarantee my business will be cited?
No, and be wary of anyone who does. Generated answers vary between users, sessions and phrasings, and no one controls model output. What I can do is measure your citation rate before and after, and demonstrate movement against a fixed prompt set and named competitors.
Do you offer AEO as a one-off or ongoing?
Both. The audit and technical implementation work well as a fixed-scope project. Entity building, content and citation monitoring are ongoing by nature, since your competitors are working on the same thing. Pricing is scoped after the consultation call, which is free.
What is llms.txt and do I need one?
It is a plain text file at the root of your site that gives AI crawlers a prioritised, clean summary of your most important content, in the same spirit as a sitemap. Adoption is still emerging and it is not yet honoured universally, but it costs very little to implement and it forces a useful exercise: deciding which pages you actually want models to draw on.