Do this small experiment with me. Open ChatGPT, or Perplexity, or trigger an AI Overview in Google, and ask a genuine question in your industry — the kind a real customer might ask before deciding to buy. Then look carefully at where the answer says it got its information.
If you are like most business owners who try this, you will feel a particular sting. The answer is good. The answer is specific. And the answer is sourced, in part, from a Quora thread written three years ago by someone with a username like "marketing_mike_88," and a Stack Exchange post where a stranger explained, with surprising patience, the exact thing your company does for a living.
Your site, meanwhile, is nowhere. You wrote the definitive guide. You have the credentials. And the machine reached past you to quote an anonymous forum reply.
It is maddening. But it is also completely logical once you understand how these systems decide whom to trust — and the good news is that the logic, once you see it, points toward a strategy you can actually use.
The numbers are not subtle
Let us start with the scale of what is happening, because it surprises people. In one 2026 analysis of more than a million AI citations across ChatGPT, Perplexity, and Google AI Overviews, community-driven platforms — Reddit and Quora chief among them — captured roughly 52.5% of all citations. More than half. The crowd, in other words, is being cited more than the entire rest of the indexed web combined.
Quora in particular punches far above its weight. Semrush's analysis found Quora to be among the most-cited platforms in Google's AI Overviews, and the pattern is even sharper in Google's AI Mode, where Quora appears around three and a half times more often than in the standard AI Overviews block.
Now, it is not a perfectly straight line upward — nothing in this field ever is. After a series of platform changes in late 2025, user-generated content briefly fell out of ChatGPT's top citation sources, dropping to a sliver of around half a percent, though there are signs of recovery. So the share moves. The platforms negotiate, the models adjust, the weather changes. But the structural point holds: AI systems have a deep, persistent appetite for community-written answers. And that appetite is not going away.
Why does the machine prefer a stranger to an expert?
This is the question that actually matters, and the answer reveals something about how AI search thinks.
When a large language model assembles an answer, it is not looking for the most beautifully designed page or the most authoritative-sounding domain. It is looking for text that directly answers a specific question, in plain language, ideally confirmed by more than one source. And forums are, structurally, a machine for producing exactly that.
Think about the shape of a Quora answer or a Stack Exchange post. It begins with a real question — phrased the way a real human would phrase it, which happens to be the way real humans phrase things to AI. Then it gives a direct answer. Then, crucially, other people pile on. They agree, they correct, they add caveats, they upvote the good replies and bury the bad ones. By the time an AI crawler arrives, the thread is a pre-digested, crowd-verified, conversationally-phrased answer to precisely the kind of question users ask.
Compare that to a typical business page. The page leads with brand language. It buries the actual answer three scrolls down, after the value proposition and the hero image. It is written to persuade, not to answer. And it has no external chorus of voices confirming that what it says is true. To a model trying to construct a trustworthy answer, the forum thread is simply easier to use and easier to believe.
There is also the matter of the question-answer match. Forums are built around questions. Your service page is built around an offer. AI search runs on questions. You see the mismatch.
What this does not mean
Before we get to the strategy, let me kill a tempting and wrong conclusion, because I have watched people leap to it.
It does not mean you should abandon your own site and pour everything into forum posting. That way lies spam, and the platforms are good at detecting it, and so are the models. A burst of thinly-disguised promotional answers will get you flagged, downranked, and possibly banned — and it will earn you nothing in AI citations, because the systems weight community content precisely because it does not read like marketing.
It also does not mean Quora and Stack Exchange are your enemies. They are, if you approach them correctly, one of the few channels where a small business can genuinely compete for AI visibility against companies a hundred times its size. The forum does not care about your ad budget. It cares whether your answer is good.
How to actually compete
So here is the strategy, and it has two halves that work together.
The first half is to make your own pages behave more like the forum content the machines love. This is within your control today. Restructure your most important pages so that each one opens by answering a real question completely — the question a customer would actually type, in their words, answered in a self-contained paragraph a model could lift whole. Add a genuine, substantive FAQ section that addresses the awkward, specific questions people really ask, not the softball ones your marketing team prefers. Write the way a knowledgeable friend would explain it over coffee, not the way a brochure would. The closer your page reads to a great forum answer, the more usable it becomes to an AI assembling a response.
The second half is to show up in the communities themselves — properly, not as a marketer in disguise. Find the questions in your field that people are genuinely asking on Quora and the relevant Stack Exchange sites, and answer them as the expert you actually are. Give the real answer, the complete one, the one that helps even if the person never becomes your customer. Mention your experience naturally where it is relevant. Do not lead with a link; let the quality of the answer earn the click, if a click comes at all. Over time, those answers accumulate, get upvoted, get crawled, and become part of the corpus the machines draw from — and when they do, your expertise is in the room even when your domain is not.
The two halves reinforce each other. Strong forum answers build your topical reputation; strong on-site answers give the machine somewhere authoritative to land when your name comes up. Neither alone is as powerful as both together.
The bigger shift to make peace with
There is a deeper change underneath all of this, and it is worth naming, because making peace with it will save you a lot of frustration.
For two decades, SEO was a game of controlling your own property. You owned the page, you optimized the page, you ranked the page. The whole discipline assumed that the thing being ranked was the thing you controlled. AI search quietly breaks that assumption. Increasingly, your visibility depends not only on what you publish on your own domain but on what the wider web — forums very much included — says about you and about your topic.
That is, in a way, a more honest world. It is harder to fake. You cannot buy your way into a thousand genuinely helpful Quora answers the way you could once buy a thousand directory links. You have to actually be useful, repeatedly, in public, where other people can confirm it.
The companies that win AI citations in the next few years will be the ones who understand that their reputation now lives in many places at once, and who show up generously in all of them. The forum is not stealing your citation. The forum is showing you where the conversation moved.
If you would like to know which questions in your niche AI is answering with someone else's words — and where your content would need to be to win those citations — that is precisely the kind of analysis Licheo runs. Contact us, and we will map it for you.