Improve your company’s visibility in AI answer engines

AI visibility describes whether and how often a company is named and cited as a source in the answers of AI answer engines such as ChatGPT, Copilot, Gemini, and Claude. Unlike a position on Google, it is not a rank but a share: out of a hundred questions asked from your market, your name shows up in a certain number of the answers. It is measured with a fixed set of queries that are repeated in the exact same wording.

In four steps. First, measure in which questions you are named today and in which you are not. Second, make your own website citable, meaning answer one question per page, put the answer up front, and back up claims with a number, a period, and a source. Third, anchor the same positioning outside your own website, in directories, comparison portals, and trade articles. Fourth, measure again every month in the same wording and derive the work list for the following month from the cited sources.

Recommended is whoever is described the same way in several sources. Answer engines check whether website, directory entries, profiles, and trade articles name the same category, the same target audience, and the same value. If these details match, a pattern emerges that the systems output as a recommendation. If they contradict each other or third-party sources are missing entirely, the mention does not happen, even with good positions in classic search.

There is no universally valid answer, because the measurement differs by market, language, and country. What can be checked is the approach. The seven criteria further up on this page can be applied to any offer: a baseline before the work, disclosed wording of the queries, a defined metric, repeated measurement, work on third-party sources, an honest statement about what cannot be guaranteed, and the handover of data and access at the end.

Specialization shows in three things: the agency measures continuously in several systems instead of once in one, it works on third-party sources and not only on your own website, and it distinguishes between GEO, AEO, and LLMO in its offer. Anyone looking for a GEO-LLMO agency or a GEO-SEO agency usually means exactly this combination: citability today, model knowledge for the coming years, and both in one hand together with classic SEO. LILOX works in this division for technical B2B companies in the DACH region and measures Austria, Germany, and Switzerland separately.

With a fixed set of queries, a fixed cadence, and a defined metric. We work with twelve real queries from your market, run them monthly on the same day in the leading AI systems, including ChatGPT, Google AI, Gemini, Claude, and also Perplexity, and report each system separately. The metric is the citation_rate, meaning citations per retrieval, plus the share of voice per query. A one-time test is not enough, because the systems’ answers fluctuate from week to week.

Through twelve real queries phrased the way your customers ask. We record whether you’re named, in what position, and from which source the mention comes, and compare that with two competitors you name. We present the result to you live in 30 minutes.

A ranking is a position in a list your customer chooses from themselves. An AI answer names three to five providers in a paragraph and already makes the shortlist. Good Google positions help, but they don’t guarantee the mention, because answer systems also draw on outside sources.

Technical points on your website often take effect within a few weeks. Work on outside sources takes longer, because articles, profiles, and directories first have to be read and processed. We think in months and show you the movement in the monthly report.

llms.txt is a text file in your website’s root directory. It shows answer systems, in short form, what you’re about and which pages are the solid sources. It doesn’t replace good page structure, but it makes finding your important content noticeably easier.

If you want to be cited, you have to be allowed to be read. That’s why we check which access your site blocks today, and deliberately open it up to the crawlers of answer systems. Many sites block them out of an old setting no one ever actually decided on.

The visibility check costs you nothing. It covers twelve real queries, comparison with two competitors you name, a technical assessment of citability, and three prioritized actions. Building the basis with an assessment per page and a topic plan runs between €3,500 and €9,900.

No. The systems decide for themselves which sources they cite, and their answers fluctuate. We guarantee the method and the measurement. Every month you see, in the same wording, where you stand, what’s changed, and which sources the mentions come from.

We measure ChatGPT, Google AI, Gemini, Claude, and also Perplexity, because that’s where most of your customers’ research lands in the DACH region. If further systems become relevant for your market, Copilot for example, we add them to the same measurement and report them separately.

SEO aims at a position in the results list, GEO and AEO aim at the answer itself. AEO, answer engine optimization, works towards your page delivering the direct answer, for example in the featured snippet. GEO, generative engine optimization, works towards ChatGPT, Copilot, or Google AI citing your page as a source when they condense several sources into one answer. The technical groundwork is the same, the goal is not.

It makes sure answer engines find your company, understand it, and cite it as a source. In practice that is four jobs: measuring which questions you are named in today and which you are not; making your own website citable, meaning answer first and evidence after; occupying third-party sources such as directories, comparison portals, and trade publications; and reporting every month, in the same wording, what has moved.

The two have to be kept apart, and we keep them apart in the report as well. A mention in an AI answer is a visibility number. Whether it turns into an inquiry shows up in the source attached to each inquiry in your CRM. We don’t promise inquiries out of ChatGPT. We show you every month where you are named and how much of it arrives at the bottom of the funnel.

Yes, and it has to be measured separately. Answer engines name different providers depending on language, location, and phrasing, so a result for Germany says little about Austria or Switzerland. That is why we measure per country, with their own questions and their own competitors. LILOX has offices in Söll in Tyrol and in Wildhaus in eastern Switzerland and works in all three markets.

Your own website is the groundwork, but it rarely suffices on its own. Answer engines prefer to back their statements with sources that don’t belong to the provider itself: provider directories, comparison portals, trade media, and industry listings. In our analyses, it is regularly these very pages that competitors get cited through. We show you which of them count for your market and occupy them one after another.