Artificial Intelligence (AI)
Artificial intelligence (AI) is the umbrella term for software that performs tasks normally requiring human judgment: parsing language, spotting patterns, drawing inferences, and producing answers. In the context of AI visibility, what matters is a narrower slice of that field '— the large language models behind assistants like ChatGPT, Gemini, and Perplexity, which read a question, pull together an answer, and in the process decide which sources and brands get named and which don't.
Why this matters for your visibility
Search used to end in a list of blue links you could click through. Increasingly it ends in a single AI-written summary that the user reads without clicking anything. Google AI Overviews alone reaches more than 2 billion people a month across 200-plus countries, and Pew Research's analysis of real browsing sessions found people click through to a traditional result in only 8% of visits when an AI summary is present, versus 15% when it isn't — roughly half the click rate. Ahrefs has measured the same pattern from the publisher side: the #1 organic result now sees a 58% lower average click-through rate when an AI Overview appears above it. If an AI system doesn't mention you, a large and growing share of the audience never reaches your site at all, no matter how well it would have ranked in the old sense.
How modern AI works
The assistants people mean by "AI" today are built on large language models: neural networks trained on huge volumes of text to predict the most likely next word, over and over, until fluent language falls out of that statistical process. Many assistants pair the model with a live retrieval step — they fetch current web pages, feed the relevant passages back into the model, and generate an answer with source attribution from that material. It's worth being precise about what this means: the model isn't comprehending your page the way a person would, it's pattern-matching over text. Content that states facts plainly and structures them clearly is easier for that process to lift and cite correctly than dense marketing copy is.
Common misunderstandings
The biggest one: that an AI answer is authoritative because it sounds confident. Models regularly produce hallucinations — fluent statements that are simply wrong — and a Columbia Journalism Review study found AI search tools misidentified basic facts about a source article (its headline, date, or URL) in more than 60% of tested queries across eight tools. A second misunderstanding is treating "AI" as one system: ChatGPT, Gemini, Perplexity, and Google AI Overviews are trained differently and cite differently, so a good showing on one says little about the others. A third, newer myth is that publishing a special llms.txt file or extra schema markup gets you preferential treatment. Google has said explicitly that no AI-specific markup or files are required for AI Overviews or AI Mode, and Google's John Mueller has confirmed no Search system reads llms.txt at all — an Ahrefs review of roughly 137,000 sites that published one found about 97% saw no measurable referral traffic from it.
Relation to AI recommendations
Whether an AI recommends you depends less on technical tricks than on what the model can find and verify about you across the web. Ahrefs' analysis of roughly 75,000 brands found that how often a brand is mentioned across the web correlates with AI citation rate at about 0.664 — roughly three times stronger than the correlation with backlinks, at about 0.218. That lines up with how these systems actually decide who to name: the 2024 SIGKDD paper that introduced the term Generative Engine Optimization (Aggarwal, Murahari, Narasimhan, Deshpande, Rajpurohit, and Kalyan) treats citation as an optimization target distinct from search ranking, and Ahrefs has separately found only 6–8% overlap between URLs ChatGPT cites and Google's top 10 for the same query. Being named more often, correctly, and favorably in AI answers is a different job from ranking, and it's the one GEO is built to measure and improve.
Example
Imagine someone asks an assistant: "Which tax advisor in Leipzig specializes in startups?" The assistant pulls in a handful of firm websites, directory listings, and blog posts, then writes three recommendations with a short reason for each. A firm whose site plainly states "we support founders through the incorporation process" and backs that with a genuinely useful guide is more likely to get named than one running only a vague slogan. The person asking may never click through at all — they might just contact the name the AI gave them. That's the moment AI now decides on your behalf.
Common questions
Is artificial intelligence the same as a search engine?
No. A search engine returns a list of links you choose from yourself. An AI assistant writes its own summarized answer and typically names only a few sources within it. Many modern assistants blend both: they search the live web and then generate an answer from what they find, but the output you see is a synthesis, not a ranked list.
Can I influence whether an AI names my brand?
Yes, mainly indirectly. The strongest documented lever is being mentioned accurately and often across other sites — that correlates with AI citation far more strongly than backlinks do. Clear, factual on-site content and a track record other people write about honestly both help. No file or markup shortcut replaces that; Google has said none is required, and llms.txt in particular has shown no measurable effect. A baseline measurement shows you where you currently stand.