Prompt and visibility audit
We build a prompt set that mirrors how your buyers actually ask, then measure how often you appear across ChatGPT, Gemini, Perplexity, Copilot and AI Overviews, and who is being named instead.
AEO / GEO
AEO and GEO Services
People are asking AI assistants the questions they used to type into Google, and getting one answer instead of ten links. We make sure your brand is inside that answer.
What are AEO and GEO, and how are they different from SEO?
Answer Engine Optimization (AEO) is optimising content so it can be extracted and served as a direct answer. Generative Engine Optimization (GEO) is the broader practice of making a brand the entity that large language models recall, trust and cite when generating a response. SEO competes for a position in a list of links. AEO and GEO compete for inclusion in a single generated answer, which depends far more on entity clarity, structured data, quotable phrasing and consistent third-party corroboration than on classic ranking factors.
50-200Prompts tracked monthly
5Assistants measured
4-8 wksFirst snippet movement
MonthlyVisibility scorecard
We buy directly on
GoogleMetaAmazon AdsTikTokYouTubeLinkedInSnapchatSpotifyMicrosoft AdsPinterestTwitchRedditGoogleMetaAmazon AdsTikTokYouTubeLinkedInSnapchatSpotifyMicrosoft AdsPinterestTwitchReddit
What is included
What you actually get
Entity optimisation
Your organisation, people, products and locations get defined once, consistently, across your site, schema, profiles and the third-party sources models read. Ambiguity is what gets you skipped.
Structured data build
Organization, Service, Product, FAQPage, HowTo, Article, Breadcrumb and Speakable markup implemented properly and validated, not pasted from a generator.
Answer-first content
Question-led headings, a direct answer in the opening 40 to 60 words, comparison tables, definitions and statistics that models can lift cleanly.
Citation and mention building
Presence on the roundups, directories, reviews, forums and industry sources that assistants disproportionately draw from when forming an opinion.
AI visibility reporting
A monthly scorecard: prompt-level share of voice, sentiment, factual accuracy, competitor citation share and assistant referral traffic.
Our process
How we run it
Map the prompts
We build 50 to 200 real buying-stage prompts for your category and baseline every assistant against them. This is the number everything else moves.
Fix the entity
Schema, naming, descriptions, profiles and knowledge panel signals get aligned so models know what you are without guessing.
Rebuild the content
Priority pages get restructured around direct answers, tables, definitions and questions. New answer pages get built for gaps competitors currently own.
Earn the corroboration
Third-party mentions, reviews and listings in the sources assistants trust, then re-measure the prompt set monthly.
Deeper
Why entity clarity decides whether you get cited
Large language models do not store your website. They store relationships between entities: what your company is, what it does, where it operates, who it serves, and how reliably different sources agree on those facts. When a user asks for a recommendation, the model reaches for the entities it has the most consistent evidence about.
That has an uncomfortable implication. A business with mediocre content but crystal clear, consistently corroborated identity often gets named ahead of a business with better content and a muddled one. Three different descriptions of what you do across your homepage, your LinkedIn and a directory listing is enough ambiguity to get skipped.
What we align, and where
- Canonical description. One sentence defining what you do, used verbatim across your site, schema, social profiles and every directory.
- Service taxonomy. Named services with their own pages, schema and internal links, rather than a single services page listing twelve things.
- Location and service area. Explicit, marked up, and consistent between your site, Google Business Profile and citations.
- People and expertise. Named authors and executives with real credentials, marked up and linked to their profiles elsewhere.
- Corroborating sources. Reviews, roundups, press, industry directories and forum mentions saying the same thing about you that you say about yourself.
The content pattern that gets extracted
Assistants lift passages, not pages. The pattern that survives extraction is consistent: a heading phrased as the question a person would actually ask, a direct answer in the first 40 to 60 words that makes sense with zero surrounding context, then the supporting detail. Tables, definitions and numbered steps get lifted far more often than flowing paragraphs. We rewrite priority pages to that shape and leave the persuasive copy where it belongs, further down.
The scoreboard
What an AEO program is actually trying to move
Every month we run the same prompt set through each assistant and record who gets named. This one chart is the reason the rest of the work exists.
- Share of prompts where your brand is named, per assistant
- Position within the answer and accuracy of the description
- Which of your pages was cited, so you know what to build more of
- Competitor citation share on the identical prompt set
- Assistant referral sessions segmented in your analytics
Share of AI answers naming each brand
Illustrative, from a 140 prompt set across ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews.
AEO vs GEO vs SEO
Three related disciplines, different scoreboards
They overlap heavily and share infrastructure. Where they differ is what counts as a win.
- We were doing the underlying work before it had an acronym
- Prompt-level measurement, not a vague AI readiness score
- Schema implemented and validated by the team that wrote the content
- Runs alongside SEO so the two feed each other instead of duplicating
| SEO | AEO | GEO | |
|---|---|---|---|
| Goal | Rank in results | Own the direct answer | Be cited in generated answers |
| Surface | Google, Bing results pages | Featured snippets, AI Overviews, voice | ChatGPT, Gemini, Perplexity, Copilot |
| Primary lever | Authority and relevance | Structure and clarity | Entity strength and corroboration |
| Key asset | Ranking pages | Schema and Q&A blocks | Consistent mentions across the web |
| Measured by | Position, sessions, conversions | Snippet and Overview capture | Prompt share of voice, citations |
| Typical horizon | 3 to 6 months | 1 to 3 months | 3 to 6 months |
Pairs well with
Services that compound this one
Ready to grow with AEO / GEO?
Get a personalised look at how our aeo / geo team can accelerate your growth.
A written read on your current setup within two business days. No obligation.
FAQ
Questions about aeo / geo
The things clients ask before they sign, answered plainly.
Does AEO or GEO actually drive revenue yet?
For some categories, clearly yes. Assistant referral traffic is small compared with classic search but converts unusually well, because the user arrives having already been recommended. The larger effect is influence: buyers who ask an assistant for a shortlist and see your name form an opinion before they ever reach a results page. We report both the measurable sessions and the citation share so you can judge for yourself.
How do you measure whether it is working?
We track a fixed prompt set every month across each assistant and report: how often you are named, in what position within the answer, whether the description is accurate, which of your pages was cited, and how your competitors' share is moving. On top of that we segment assistant referral traffic in analytics.
Can you guarantee ChatGPT will recommend my business?
No. Model outputs are probabilistic and change with every training and retrieval update. What we can do is control the inputs those systems read, measure your citation share honestly, and move it. Anyone promising a guaranteed AI recommendation is selling you something they cannot deliver.
What is the difference between AEO and GEO in practice?
In practice they share about 70% of the same work. AEO leans toward structure, formatting a page so an answer can be extracted from it. GEO leans toward reputation, making sure the model has enough consistent evidence about your brand across the web to name you confidently. We sell them together because doing one without the other tends to stall.
Do I need to rewrite my whole website?
Almost never. Most programs start by restructuring 20 to 40 existing high-intent pages, adding schema sitewide, and building a small set of new answer pages for gaps. Full rewrites are a last resort, usually when the site has no crawlable content at all.
How long before AI assistants start citing us?
Structured data and answer formatting can produce featured snippet and AI Overview movement in 4 to 8 weeks. Getting consistently named inside conversational assistants typically takes 3 to 6 months, because it depends on third-party corroboration accumulating and on model retrieval refreshing.
Does AEO help with voice search too?
Yes. Voice assistants read from the same extracted-answer infrastructure, so the Speakable markup, concise answer blocks and FAQ schema we build for AEO carry over directly.
Next step
Ready to see what your budget should actually be doing?
Send us your current campaigns. We will show you, line by line, where the waste is and what we would do differently. No pitch deck, no obligation.