AI visibility: getting cited by AI answer engines

What GEO is and why it matters

GEO stands for Generative Engine Optimization. Where classic SEO works to rank your pages in a list of blue links, GEO is about getting your site mentioned and cited inside the answers that AI engines write. When someone asks ChatGPT, Gemini, or Perplexity a question in your niche, the goal is for your brand and your URLs to appear in the response.

This matters because more and more searches end inside an AI answer, without a click to a results page. If you are invisible to the engines, you lose that traffic and that trust signal. AI visibility tells you, in plain numbers, whether the engines already know and recommend you, or whether your competitors own the conversation.

How Lume's multi-engine check works

Lume runs your keywords against three answer engines at once: Gemini, ChatGPT, and Perplexity. For each keyword it asks the engine a real question and reads the answer back, looking for whether your site is mentioned or cited. Perplexity searches the live web as it answers, so it reflects what is findable online right now, while the others draw more on what the model already knows.

Because the three engines behave differently, checking all of them gives a more honest picture than any single one. A keyword where ChatGPT ignores you but Perplexity cites you tells you the gap is awareness, not findability — a useful distinction when you decide what to fix.

Reading mention rate and Share of Voice

Two headline numbers anchor the report. The per-engine mention rate is the share of your keywords for which a given engine mentioned or cited you — read it engine by engine to see where you are strong and where you are absent. Cross-engine Share of Voice looks across all the answers and counts which domains the AIs cite most often; your own domain is flagged so you can spot it instantly against the field.

• Mention rate: how often each engine names or cites you, per keyword — your raw visibility on that engine.
• Share of Voice: which domains the AIs cite most across all answers — your slice of the cited-source pie versus competitors.

Competitor benchmark, tracked prompts, and AI Overviews

From the same set of answers, Lume derives a you-vs-competitors benchmark at no extra cost: it shows how your visibility stacks up against the other domains the engines cite for your keywords, so you can see at a glance whether you lead or trail. Because it reuses answers Lume already collected, the benchmark is free and stays in sync with your latest check.

If you have specific questions you care about, register them as tracked custom prompts — the exact phrasing your customers would type. Lume runs those prompts on every check so you watch the questions that matter to your business, not just generic keywords. Separately, Google AI Overviews occupancy tells you whether an AI Overview shows for a query at all, and if it does, whether your site is among the sources it cites.

• Competitor benchmark: free, derived from the same answers — your visibility ranked against the domains the engines cite.
• Tracked prompts: register the exact questions you care about and watch them on every check.
• AI Overviews occupancy: does an AI Overview appear for the query, and are you cited in it?

Brand sentiment and tracking over time

Beyond whether you are mentioned, brand sentiment captures how the LLMs talk about your brand — whether the language around your name skews positive, neutral, or negative. This surfaces reputation problems that a pure visibility score would miss: being cited often but described unfavorably is a different problem from not being cited at all.

Every check is saved to your history, so you can run AI visibility weekly and watch the trend rather than read a single snapshot. Tracking over time is where the value compounds: you can tie a content or PR change to a rise in mention rate, catch a competitor pulling ahead in Share of Voice, or confirm that a sentiment dip recovered after you addressed it.

Tips for acting on the results

The report is a to-do list, not just a scorecard. Start with the keywords where every engine ignores you — those are the clearest content gaps. Then look at keywords where Perplexity cites you but Gemini and ChatGPT do not: that pattern points to a page that exists and is findable, but is not yet authoritative enough for the models to recommend from memory.

• Prioritize keywords where you score zero across all three engines — the biggest, clearest gaps.
• Study the domains that top your Share of Voice: the pages the AIs already cite show what good looks like.
• Re-run weekly after each change so history can attribute movement to what you did.