How ChatGPT, Gemini, Perplexity and Copilot Decide Who to Mention
Ask the same question of ChatGPT, Gemini, Perplexity, and Microsoft Copilot and you will often get four different answers, citing four different sets of sources. That is not a quirk. Each engine builds its answer differently, from different places, which means being named by one is no guarantee of being named by another.
Here is how each of the four biggest AI engines decides who to mention, and what that means for getting your brand into their answers. For the full reference on each platform, see our AI Platforms guide.
They all start from two things
Under the hood, every one of these engines does some mix of two things. It draws on what the model already learned during training, which is roughly what it knows about you, and it retrieves fresh information from the live web or a search index, which is what it finds about you right now. The balance between those two, and which index it reaches into, is where the engines differ. If that split is new to you, we cover it in Memory versus Live in our glossary.
ChatGPT
ChatGPT answers from a mix of its training knowledge and, when it searches, live results from its own web search system. Its citations lean heavily on a small set of highly trusted sources, Wikipedia above all. One large analysis found Wikipedia made up nearly half of ChatGPT's most-cited sources.
To improve your chances of being mentioned, make sure your brand is present and accurate in the sources ChatGPT trusts. Start with a solid Wikipedia and Wikidata presence if you qualify, keep your site crawlable to OpenAI's crawlers, and structure your pages so a clean answer can be lifted out of them.
Gemini
Gemini is Google's model, and it draws on Google's understanding of the web. It is closely tied to Google's own AI search surfaces, AI Overviews and AI Mode, which are built on the same foundations. Google-owned YouTube features prominently in those surfaces, and classic Google ranking signals still carry weight.
To improve your chances, the strongest lever is the SEO you already know. Rank well in Google, since much of what Gemini and Google's AI surfaces cite comes from the top results. Add a presence on YouTube for explainer and how-to topics, and make sure you allow Google's AI crawler (Google-Extended) to access your site.
Perplexity
Perplexity was built from the start as an answer engine. It searches the live web for almost every query and puts citations front and center, which makes it the most retrieval-driven of the four. It has leaned heavily on community sources like Reddit, though that share has been falling, and it strongly favors fresh content.
To improve your chances, keep your key pages current, since freshness matters more here than almost anywhere else. Make sure PerplexityBot can crawl you, and build an honest presence in the community sources and cited pages that come up for your category.
Microsoft Copilot
Copilot is grounded in Bing's index. In practice, that means its answers largely reflect what is crawlable and ranking well in Bing, not Google. This is the one many brands overlook, because they optimize for Google and forget Bing entirely.
To improve your chances, treat Bing as the lever. Submit your site in Bing Webmaster Tools, fix anything blocking Bingbot, and work on your Bing rankings specifically. If you are invisible in Bing, you will be invisible in Copilot.
The common thread, and the catch
For all their differences, the engines share a foundation. Every one of them rewards a site that is crawlable, content that is fresh and clearly structured, and a brand that is present and accurate across the trusted sources in its category. Get those fundamentals right and you improve your odds everywhere at once.
The catch is that the specifics diverge, and sometimes sharply. Wikipedia can make or break you on ChatGPT while barely registering elsewhere. Google rank drives Gemini. Freshness and community presence drive Perplexity. Bing drives Copilot. Because of that, you cannot assume that winning one engine means winning the others. You have to check each one.
That is exactly what Diploria does. It tracks how you show up across all of these engines, plus Grok and Google's two AI search surfaces, so you can see where you stand on each rather than guessing. The Competitive Intelligence use case digs into how you compare per engine.