Mention rate
How often AI assistants mention you at all when given a tracked prompt. The first hurdle, and the one most businesses haven't cleared.
Answer Engine Optimisation
Answer Engine Optimisation is the work of making your business the answer when a customer asks an AI assistant who to buy from. Not a link on a page of links: the name in the reply, with a reason.
Answer Engine Optimisation (AEO) is the practice of strengthening the signals AI assistants rely on when they decide which businesses to name, so that when a buyer asks ChatGPT, Google's AI Mode or AI Overviews, Gemini or Claude a question with money behind it, your business is mentioned, recommended and described correctly.
It covers four things an AI assistant needs before it will put your name in an answer: it has to be able to read you, understand you, trust you, and find you again next month. AEO is the work of making all four true, and measuring them.
Buyers moved first. The AI assistants answer the question and hand over a shortlist; the shortlist is where the sale is won or lost.
94% used AI somewhere in their most recent purchase.
95% of winners were already on the buyer's day-one shortlist.
51% now start their purchasing research inside AI tools.
Same site, different game. SEO earned you a position in a list; AEO earns you a place in a sentence.
| The difference | SEO | AEO |
|---|---|---|
| Where the decision happens | On your site, after a click | Inside the answer, often before any click |
| The unit of work | Keywords and pages | Buyer prompts and the facts that answer them |
| What gets you chosen | Relevance and authority, mostly links | Clear identity, consistent details, direct answers, independent proof |
| Who reads your site | A crawler that indexes | A model that summarises, quotes and compares, plus its crawlers |
| What winning looks like | Position one, more traffic | Named, recommended, described correctly |
| How it is measured | Rankings, impressions, clicks | Mention rate, recommendation rate, position, per platform |
| How fast it changes | Slowly, with algorithm updates | Monthly: new models, new features, new competitors in the answer |
There is no ranking algorithm to game. An AI assistant assembles an answer from what it can read and what it can verify. Four things decide whether that includes you.
Crawlers allowed in, pages that render as text, facts on the page rather than behind a script, and a site fast enough to be fetched in the moment. If the AI assistant can't read it, it doesn't exist.
A clear statement of what you do, for whom, where, and at what price, repeated identically on your site, your listings and your registers. Pages that answer the buyer's question in the first paragraph, not the fifth.
AI assistants are cautious recommenders. Reviews on platforms they cite, directories and registers that agree with your site, press and mentions from sources they already trust. Proof that travels beyond your own domain.
Models get updated, competitors publish, and platforms add features. A business that was being recommended in March can be missing by June without having changed a thing. AEO includes watching, and putting it right.
Google's AI Overviews and AI Mode reach the most people because they sit inside Search. ChatGPT is the AI assistant people name when they say "I asked AI". Gemini reaches Google's own users, and Claude is the assistant a growing share of professionals and businesses use. Each one reads slightly different sources and behaves slightly differently, so AEO is done per platform and measured per platform.
What they have in common: none of them can be tricked for long. Prompt injection, fake reviews and link schemes get closed quickly and leave a mark. The durable route is the boring one: be easy to read, be consistent, be provably good, and keep it that way.
There are seven strands of work, and they are done together because each one feeds the others. This is the programme Recrawled runs; what we deliver has the detail.
Classic SEO hygiene comes with all of it, because AI assistants and search engines read the same site.
We track three numbers for each prompt on each platform, and we do it every month. Behind them sit the numbers that pay for the work: enquiries, sales and bookings.
How often AI assistants mention you at all when given a tracked prompt. The first hurdle, and the one most businesses haven't cleared.
How often they actively recommend you, with a reason, rather than listing you among others. This is the number that moves enquiries.
Where you sit when several names are given. Buyers rarely go past the first two or three.
Broadly, yes. Answer Engine Optimisation, Generative Engine Optimisation and LLM optimisation all describe the same job: making a business visible and recommended inside AI assistants. We say AEO because the thing being optimised for is the answer, and that keeps everyone focused on what the customer actually sees.
No. AI assistants and search engines read the same site, so classic SEO hygiene is part of the work and comes with it. What changes is the goal: being named and recommended in an answer, rather than ranked in a list of links that fewer people click.
No, though it overlaps. Digital PR earns new coverage by pitching stories to journalists, and we don't do that. What we do is make the evidence that already exists about you accurate, consistent and complete, because that's what an AI assistant checks before it recommends anyone: your reviews, your listings, the directories and registers you appear in, and any coverage you already have, all saying what your site says. If you have a PR agency, we plan around them, and their coverage works harder once the rest of the picture agrees with it.
First mentions often appear within the first two or three months, once the site can be read and the facts agree everywhere. Recommendations take longer because they depend on independent proof, which takes time to build. That's why the programme runs for six months, or twelve on Steady, and is measured every month; our work shows what the curve has looked like.
Some of it, and our research explains how. Fixing crawlability, making details consistent and answering buyer prompts directly are all things a capable team can do. The parts that take sustained effort are the monthly measurement across platforms, the independent proof, and keeping up as the platforms change. A free assessment shows you where you stand either way.
Yes, and often fastest. A local or specialist business has a small set of questions to win, and AI assistants reward clear identity and consistent details more than size. See AEO for B2C and local businesses.
For twenty years or more each, alongside digital marketing, sales and web development. The classic work is familiar ground, and the hygiene AEO depends on is part of the programme anyway. What changed is where the decision happens: a buyer who used to compare ten blue links now gets two or three names and a reason. We built for that rather than adding it to a service designed for something else, and we work alongside an existing agency rather than over it. Where we came from.
What replaces clicks if the answer happens inside the AI assistant.
ResearchHow to choose the exact questions you need to be winning in.
ResearchTurning a company website into a cite-able answer surface.
Tell us about you and your business, and we'll kick things off immediately.