One study, seven AI engines, very little overlap – what it means for the posts you write next. By the Blogawesome team.
If you are a founder or a one-person marketing team, you probably think of "showing up in AI" as one job. Ask ChatGPT, get mentioned, done. A study published this week suggests it is closer to seven different jobs – and that changes how AI content marketing should be planned.
Seven engines, one question, different answers
Conduit, an AI agent company in hospitality, pulled 1,454 citations from ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Microsoft Copilot and Claude. Each engine answered the same 33 software-buying questions.
The headline finding: of 810 unique pages cited, 614 – 76% – were cited by only one engine. Only two pages were cited by all seven. Even ChatGPT and Claude, both chat assistants, were the most different pair in the study.
The study covers one industry, hospitality software, so treat the exact numbers as a signal rather than a rule for your market. But the direction is hard to ignore: being visible in one AI engine tells you little about the others.
Why this matters for AI content marketing
Most small teams check one assistant, see their brand (or a competitor) named, and draw a conclusion. If engines cite mostly different pages, that single check can be misleading in both directions. You might be invisible in Perplexity while doing fine in ChatGPT, or the reverse.
It also means a single "AI-optimised" article is unlikely to cover you everywhere. What gets you named is a spread of pages, each useful to a different kind of answer. If you are new to the idea, our explainer on what AI brand visibility is covers the basics.
What each engine seemed to reward
In Conduit's data, the engines leaned on different types of page:
ChatGPT cited product and feature pages most – 40% of its citations.
Claude, Copilot and Perplexity leaned on "best X" listicles – between 50% and 68% of their citations.
Perplexity cited about 10 sources per answer, roughly two and a half times Gemini's 3.6.
Claude gave no citations at all on 8 of its 33 answers, mostly "how do I" process questions.
For a small team, the practical read is simple. Clear product pages help in one place. Honest comparison and roundup content helps in others. Engines that cite more sources give you more chances to be one of them.
Google is adding more links, not fewer
On 15 October, Google announced updates to AI Mode and AI Overviews: more inline links shown higher in responses, a test of a dedicated section for official pages, and more prominent links to related articles at the end of answers.
That is good news for smaller brands. More link slots inside AI answers means more room for a well-researched page from a company that does not have a big domain behind it – if the page clearly answers the question.
A simple plan for small teams
Make your product pages answerable. Say plainly what you do, who it is for and what it replaces. These are the pages ChatGPT reached for most in the study.
Publish honest comparison content. "Best tools for X" and "X vs Y" posts that include you alongside real alternatives match what several engines cite.
Check every engine, not one. Run the same buying questions in ChatGPT, Gemini, Perplexity and Claude and note who gets named instead of you.
Write to the gaps. Where a competitor is named and you are not, that keyword is your next post.
Re-check after you publish. AI answers shift. A monthly check per engine shows whether new posts are moving anything.
A note from us
This is the exact loop Blogawesome is built for. It checks whether ChatGPT, Gemini, Perplexity and Claude recommend your brand for the keywords you care about, shows the gaps and who AI names instead, then researches, drafts and publishes on-brand posts in your voice to WordPress, Shopify or a webhook. We run it on our own blog too.
Want to see where you stand across the engines? Get your free AI visibility audit.