Get cited in the answer, not ranked below it.
Buyers increasingly get their answer without ever seeing a results page. If the answer does not mention you, you were never in the running.
- You want visibility inside AI answers, ChatGPT, Gemini, Perplexity, Copilot.
- You want this measured, not asserted.
- You want the technical work shipped, not recommended in a deck.
What the free Growth Audit gives you
- Improvement areas identified from your own data, not from a template.
- Hidden potential in your current setup, surfaced and quantified.
- Practical, ROI-focused recommendations you can act on next week.
Done with the uncertainty about the future of search?
The honest position: nobody knows how fast this shifts. What is already true is that answers are being generated above the links, and somebody is being cited in them.
- Competitors appear in AI answers for your category while you do not.
- Less control over the buyer journey, as decisions happen out of sight.
- No way to tell whether any of it is sending you traffic.
All three are measurable. That is where we start, because everything after it is guesswork otherwise.
How we run GEO
We baseline where you and your competitors are cited today, trace which sources the engines pull from, ship the technical and editorial work, then report citation share per engine every month.
The authority signals overlap heavily with ordinary SEO, topical depth, entity clarity, structured data, external citations. The surface and the scorecard are what differ. If your conventional SEO is weak this will be weak too; they compound rather than replace each other.
What GEO actually means
GEO, AEO and “AI SEO” describe the same work from different angles: optimising for engines that answer rather than list. Here is the plain version, including the parts that are not settled.
GEO, AEO and AI SEO are one job
Generative Engine Optimisation, Answer Engine Optimisation and AI SEO are competing labels for a single discipline. Anyone selling them as three services is selling you the same work three times.
Four engines, four different habits
AI Overviews lean on sources that already win featured snippets and on schema. Perplexity favours fresh, clearly structured pages. ChatGPT mixes training data with live browsing and rewards factual density. Copilot leans on Bing’s index and on timeliness. One tactic does not move all four, so we report them separately rather than as a single score.
A mention and a citation are not the same thing
A mention names you in the answer with no link. A citation links back. Both are worth having and they behave differently, mentions build recognition, citations also send traffic, so we report them separately. Most tools blend the two into one number, which flatters the result.
Ranking first, citation second
Content that already ranks on page one is substantially more likely to be cited: the engines lean on the same authority signals, and several of them retrieve live from the results. That makes this a sequence, not a choice, which is why we will tell you to fix conventional SEO first if it is weak.
Check you are not blocking them first
A large share of sites block AI crawlers, frequently by accident, a robots.txt rule added years ago, or a bot-protection default. It is the first thing we check, because everything else is wasted if GPTBot cannot read the page. It is also an opportunity: where competitors block, the field thins.
llms.txt: cheap, unproven, ship it anyway
A plain text file at your root telling engines what matters. It is not a confirmed ranking factor anywhere and we will not pretend otherwise, but the spec is stable, some engines read it, and it costs one pull request.
Four things we report, not one
“AI visibility” collapses into a single vanity number very easily. It is really four separate questions, and a brand can be strong on one and absent on another.
Citation frequency
How often each engine cites or mentions you across a fixed, repeatable prompt set. Tracked per engine, because a gain in Perplexity tells you nothing about ChatGPT.
Answers: Are we in the answer at all?Share of voice
Your citations as a proportion of all brands cited for the same prompts. The AI equivalent of market share, and the number that shows whether you are gaining or the category is simply growing.
Answers: Are we in it more than our competitors?Sentiment
How you are characterised when you do appear. A brand named only in the context of complaints has visibility of the wrong kind, and that will not show up in a citation count.
Answers: Is being mentioned actually helping us?Query coverage
The share of relevant prompts where you appear at all. Brands routinely own “what is X” and are entirely absent from “best X for Y”, which is where the buying happens.
Answers: Which questions are we missing entirely?How a GEO engagement runs
Five phases. Each produces something you keep, whether or not you continue with us.
Citation baseline
Weeks 1–2We script buyer-intent prompts for your category and log which brands each engine cites. You see your citation share, your competitors’, and the prompts where nobody has been cited yet, usually the fastest wins available.
What you keep
- A per-engine baseline you can re-run later
- Competitor citation share on the same prompts
- Uncontested prompts, ranked by commercial value
Source mapping
Weeks 2–3For prompts where your category gets cited, we trace what the engines actually pull from, reference sites, forums, trade publications, competitor pages. You cannot earn a citation without knowing what is being cited instead.
What you keep
- A source map per prompt cluster
- Which sources you can realistically influence, and which you cannot
- Where a placement is worth pursuing, and where it is not
Access, entities and structure
Weeks 3–12First, confirm the engines can actually read you, robots.txt, bot protection, and server logs showing which AI crawlers reach which pages. Then the entity work that makes a brand recognisable to a model rather than merely present on a page. Delivered as changes your developers review, not a list of suggestions.
What you keep
- Crawler access verified, and AI bot traffic read from your server logs
- Entity signals: consistent naming, organisation schema, Wikidata where you qualify
- Structured data implemented and validated, not merely present
- llms.txt, with an honest note on its current limits
Content worth citing
OngoingEngines do not read a page, they lift pieces out of it. So structure decides as much as substance: a direct answer in the opening lines, sections that stand alone without the one above them, and tables and steps that survive being extracted. On top of that you need something worth quoting, original data, a genuinely definitive explainer, or a position only you can hold. Volume does not work here.
What you keep
- Briefs written against prompts your buyers actually run
- Answer-first openings, so the extractable part is the part you want extracted
- Sections rewritten to stand alone, plus comparison tables and step lists
- Original research where you have data nobody else has
Monthly citation scorecard
OngoingCitation share per engine, movement since last month, and AI-assistant referral traffic from your analytics, so the work ties to sessions and revenue rather than to screenshots.
What you keep
- Citation share per engine, tracked over time
- Prompt-level wins and losses
- AI-assistant referrals, beside every other channel
We measure this on our own traffic first
Our analytics classify visits arriving from AI assistants as a distinct channel, including referrers that disguise themselves, eleven detection cases, two of them spoofed hosts. It sits in the same reporting as paid, organic and email, so an AI-assistant session is comparable with every other session rather than living on a separate slide. We built that here before recommending it to anyone, and we will show you the configuration.
Where we fit, and where we do not
Several kinds of supplier touch this. Here is the comparison as we would make it if you were not in the room.
| A conventional SEO agency | A prompt-monitoring tool | Lockbox | |
|---|---|---|---|
| Measurement | Reports rankings. AI visibility is usually absent or anecdotal. | Shows citations across engines, often well. That is the entire product. | Citation share per engine, joined to AI-assistant referral traffic and revenue in the same reporting as every other channel. |
| Implementation | Strong on technical SEO. Entity and llms.txt work is often new to them. | None. A tool cannot ship anything to your site. | We implement it, structured data, entity markup, llms.txt, content restructuring. |
| Content | Briefed against keyword volume, which is the wrong target here. | Not offered. | Briefed against the prompts buyers actually run, and against whatever is being cited instead of you. |
| Crawler access | May check robots.txt. Server-log analysis of AI bots is rarely offered. | Cannot see your logs at all. It only queries the engines from outside. | We read your server logs to see which AI crawlers actually reach which pages, and fix the access problems that make everything else pointless. |
| Honest limits | Rarely says the category is still unproven. | Will not tell you a citation produced no revenue. | We will tell you when your conventional SEO or your tracking should be fixed first, and when a prompt is not worth chasing. |
What you walk away with
No obligation, no pitch deck. If we can't see a way to help, we'll tell you that instead.
- Improvement areas identified from your own data, not from a template.
- Hidden potential in your current setup, surfaced and quantified.
- Practical, ROI-focused recommendations you can act on next week.
Questions we get asked
Is GEO different enough from SEO to pay for separately?
How do you measure something inside a chatbot?
Should we deploy llms.txt?
Does AI search replace organic search?
What is the difference between a mention and a citation?
Could we be blocking the AI crawlers without knowing?
How long before we get cited?
Find out what's actually holding your growth back
A free Growth Audit, built on your own data. You'll get the bottleneck and the first move, whether or not you work with us.