Everything you'd ask
before you sign up.
What GEO is, which engines we track, how the data works, what it costs, and how your clients' data is kept separate. If it's not here, the contact form reaches a human.
Does generative engine optimisation actually work?
It can, when it is measured. AI engines lean on clear facts, trusted third party pages and consistent details across the web. Improving those things tends to raise how often a brand is named, but no one can promise a mention, so it helps to track the same questions over time and judge by the trend.
Is SEO dead now that people use AI?
No. AI engines still read the web, and many of them use search results to find pages. Good SEO is still the base. GEO adds work on facts, citations, reviews and question-led content so the AI can use your pages with confidence.
What is GEO, Generative Engine Optimisation?
GEO is the practice of getting your brand named, cited and recommended inside AI answers, the summaries ChatGPT, Gemini, Perplexity and Google's AI Overviews return instead of ten blue links. Where SEO optimises for a ranking position, GEO optimises for being the answer.
How is this different from normal SEO tools?
SEO suites measure where you rank on a results page. Axiom GEO measures whether the AI mentions you at all, where you land in its answer, which competitors it names instead, and the actual sources it drew on. We also include the classic rank-tracking / GA4 / Search Console stack, so both live in one login.
How does Axiom GEO collect AI answers?
Mostly through each engine's API, on the schedule set for each prompt, from daily to monthly. ChatGPT, Perplexity and Gemini are asked with web search on, so their answers include cited pages. Claude is asked without web search, which shows what the model itself believes, and can also be asked with web search as a separate engine. Google AI Mode, and ChatGPT and Gemini as their apps show them, are read through a data provider. Consumer apps can still differ a little because of personal history and location.
Can it tell me when an AI says something wrong about my business?
Yes, on Professional and above. You approve a fact sheet, and every answer is broken into claims and checked against it. Wrong or out of date claims are grouped into issues, with the page the engine used or, where no source is named, the pages that state the same thing.
Which AI engines do you track?
ChatGPT, Claude, Gemini, Perplexity, Google AI Mode and Google AI Overviews, with ChatGPT and Gemini also available as their apps show them. Your prompts run on the schedule you set for each one, from daily to monthly, and every result is read for mention, position, sentiment, competitors and the sources cited. Microsoft Copilot citations come from Bing Webmaster Tools uploads.
Where do the 'citation sources' come from?
The engines that search the web (ChatGPT, Perplexity, Gemini and Google AI Mode) return the pages they cite. We classify each one, directory, review site, listicle, social, a company's own page, and let you open the real URLs, so you can see exactly what the AI learned from.
Can you track different countries or regions?
Yes. Run the same prompts per market and compare Share of Voice by region, useful when a brand is strong in one country and invisible in another.
How does pricing work?
A monthly plan fee that includes a credit allowance. Starter is £500, Professional £799 and Premium £1,499 a month, with custom Enterprise plans. One credit is 1p, every action has a published credit price, and every credit spent is logged.
Can I add more clients later?
Yes. Starter covers one client, Professional up to three and Premium up to ten, with Enterprise for more. Move up a plan when you need more workspaces or credits.
Can I cancel any time?
Yes, from the billing portal. Service continues to the end of the paid period; there are no long-term contracts.
Is each client's data kept separate?
Yes. Each workspace is kept separate, and every query is limited to the workspace it belongs to. OAuth tokens, credentials and analytics never mix between clients, and sub-processors never see one client's data mingled with another's.
Who can see the classifier / admin tools?
Operator tooling, like the citation classifier worklist, is gated to your admin users and never rendered to the client viewing their own dashboard.
Still have a question?
The contact form reaches a human, no bot, no queue.