We Tracked 12 Months of AI Search Citations. Here’s Who’s Winning and Why.

About a year ago, we started paying attention to something most agencies were ignoring.

Every time we ran a search query through ChatGPT, Perplexity, Google’s AI Overviews, or Gemini, we started noting which brands showed up in the citations. Not which brands ranked on page one of traditional search. Which brands the AI actually named, quoted, and linked to when it synthesized an answer. We tracked it across client industries, across platforms, and across query types. We cross-referenced what we were seeing with every major citation study that came out during that period.

Twelve months later, I can tell you this: the brands winning in AI search look nothing like the brands that dominated traditional SEO. The playbook is different. The signals are different. And most businesses have no idea this shift is even happening.

Here’s what we found.

The Citation Landscape Has Exploded

Let me put some numbers on the table so you understand the scale of what we’re talking about.

AI Overviews now appear in roughly 50% of Google search queries in the United States. For certain categories like health, finance, and technology, that number is significantly higher. When AI Overviews appear, the zero-click rate jumps to approximately 83%. That means 8 out of 10 people get their answer directly from the AI summary and never click through to any website.

Google’s newer AI Mode is even more aggressive. Early data shows around 93% of AI Mode searches end without a click. That’s not a typo. Ninety-three percent.

Meanwhile, ChatGPT Search is growing fast. Perplexity is growing fast. Gemini is growing fast. These platforms don’t show ten blue links. They synthesize answers from multiple sources and cite them inline. The brands that get cited get visibility. The brands that don’t get cited don’t exist in these conversations.

When we started tracking this a year ago, the conventional wisdom was that AI search was a novelty. Something to keep an eye on. Twelve months later, it’s eating traditional search alive. Ars Technica reported a roughly 40% average decline in click-through rates on informational and research-intent keywords once AI Overview summaries started appearing. That’s not a dip. That’s a structural shift. The traffic that used to flow through traditional search results is being absorbed by AI-generated answers, and the only brands that benefit are the ones being cited in those answers.

The businesses that figured this out early are pulling away from everyone else.

Reddit Is Running the Table

This was the single most surprising finding across everything we tracked.

Reddit is the number one most-cited source across every major AI search engine. Not just one of them. All of them. ChatGPT, Perplexity, Gemini, Google AI Overviews. Reddit shows up everywhere.

The numbers are staggering. Reddit appears in roughly 40% of AI-generated responses across major platforms. Perplexity cites Reddit in nearly 47% of its responses. ChatGPT has Reddit in its top three sources consistently. Even Google’s own AI Overviews lean heavily on Reddit threads for certain query types.

Why? Because AI systems are trained to value authentic human perspective. Reddit threads are full of real people sharing real experiences, recommendations, and opinions. They’re messy, unfiltered, and often contradictory. But they represent genuine human signal in a way that polished corporate blog posts don’t.

Here’s what this means for your business: the conversations happening about your brand on Reddit matter more than ever. If someone asks “best digital marketing agency in Tulsa” on Reddit and your name comes up in the responses, that thread has a real chance of being cited by AI systems when someone asks a similar question through ChatGPT or Perplexity. We’ve seen this with our own clients. A single well-received Reddit mention in a relevant subreddit generated more AI visibility than months of traditional SEO content.

If you’re not monitoring what’s being said about your brand on Reddit, and you’re not participating authentically in the communities where your customers hang out, you’re invisible to the fastest-growing search platforms on the planet.

And I want to be clear about what “participating authentically” means. It doesn’t mean creating a brand account and posting promotional content in subreddits. Reddit communities will destroy you for that. It means genuinely contributing knowledge and perspective in the spaces where your expertise is relevant. Answering questions. Sharing experiences. Being a real person who happens to be good at what they do. That’s what AI systems pick up on. That’s what gets cited. The moment you try to game it with corporate speak and thinly veiled self-promotion, the community rejects you and the AI systems never see you.

LinkedIn Became the Dark Horse

This one changed how we think about content strategy for ourselves and every client we work with.

LinkedIn went from being the 11th most-cited domain in ChatGPT in November 2025 to the 5th most-cited domain by February 2026. That’s the largest jump in citation authority any platform experienced during our tracking period. For professional and B2B queries specifically, LinkedIn is now the number one most-cited source across ChatGPT, Gemini, Google AI Overviews, AI Mode, Copilot, and Perplexity.

Let me say that differently: if someone asks an AI system a question about marketing, consulting, finance, leadership, technology, or any professional topic, LinkedIn content is more likely to be cited than any other source on the internet. More than Wikipedia. More than Forbes. More than any industry publication.

And here’s the kicker: 95% of cited LinkedIn content is original posts and articles. Not reshares. Not reposts. Original content from people who are consistently publishing on the platform. AI systems favor primary perspectives. They want the source, not the echo.

We saw this play out directly. The LinkedIn posts I’ve been writing about marketing analytics, AI search, and data-driven strategy aren’t just generating engagement on the platform. They’re building entity authority that feeds back into how AI systems evaluate TMG as a source. Every post is a signal. Every article is a data point that tells ChatGPT, Perplexity, and Google’s AI that Dominick Montgomery and The Montgomery Group have something credible to say about digital marketing.

The takeaway is simple but most people aren’t acting on it: your LinkedIn content strategy is now your AI search strategy. They’re the same thing. The businesses that figured this out six months ago are already showing up in AI citations. The businesses that figure it out six months from now will be playing catch-up for years.

Brand Mentions Are the New Backlinks

For twenty years, backlinks were the currency of SEO. You wanted links from authoritative sites pointing to your content because Google used them as a trust signal. The more quality links you had, the higher you ranked.

AI search doesn’t work this way.

What we tracked across 12 months is a clear shift from link authority to mention authority. Brands that are frequently mentioned across authoritative sources, even without a hyperlink, are getting cited by AI systems at significantly higher rates than brands with massive backlink profiles but low mention frequency.

The research backs this up. Studies in 2026 show that brands with fewer links but higher mention density on social platforms and news sites are often cited above legacy brands with massive backlink profiles. One study found that a minimum of 5 to 7 independent authoritative mentions are needed to establish enough “entity confidence” for an AI system to begin citing you as a primary source. Another found that adding citations from authoritative sources can increase AI visibility by 25 to 99%.

Think about what this means for your strategy. Traditional link building campaigns, where you guest post on random blogs to get a backlink, are losing relevance. What matters now is whether real people and real publications are talking about your brand in context. A mention in a relevant Reddit thread. A quote in an industry article. A reference in a LinkedIn post from someone in your space. A podcast interview where your name comes up naturally.

We shifted our own approach based on this. Instead of chasing backlinks for TMG, we focused on being present in conversations where our expertise is relevant. Publishing consistently on LinkedIn. Contributing genuine insights in forums and communities. Creating content worth referencing. The result: our entity authority is building faster than any link campaign could deliver.

I’ll put it bluntly: if your current SEO strategy is still primarily focused on link building, you’re investing in a depreciating asset. Links still matter for traditional Google rankings. Nobody’s saying throw them out completely. But in the AI citation economy, mention density is the leading indicator. And every dollar you spend chasing a backlink on a random DA 40 blog is a dollar you could have spent building the kind of brand presence that gets you cited across every AI platform simultaneously.

What Content Gets Cited (And What Gets Ignored)

Not all content is created equal in the AI citation economy. We tracked which types of content consistently showed up in AI responses and which types were invisible despite performing well in traditional search.

Here’s what we found.

Listicles dominate. Across AI Mode, ChatGPT, and Perplexity, listicles account for roughly 22% of all citations. Articles account for about 17%, and product pages about 14%. AI systems love structured, scannable content because it’s easy to extract specific answers from.

Original data and analysis win. Content that includes original calculations, frameworks, or data points gets cited at disproportionate rates. This makes sense. AI systems are looking for definitive answers. If your content contains a proprietary framework or a specific calculation that nobody else has published, the AI has to cite you because it can’t get that information anywhere else.

Schema markup is a multiplier. Sites with properly implemented structured data schema are getting cited roughly 3.2 times more often than sites without it, according to one analysis. FAQPage schema and Article schema with author information have the strongest correlation with citation rates. It’s not that schema forces a citation. It’s that schema makes your content machine-readable, and AI systems are machines.

Author credibility signals matter. AI systems are increasingly evaluating who wrote the content, not just what’s on the page. Content with clear author attribution, author bios, and verifiable credentials gets cited more frequently. This is consistent with what Google’s own E-E-A-T framework has been pushing for years, but AI systems take it further because they’re actively trying to determine whether a source is trustworthy enough to cite in a synthesized answer.

ChatGPT pulls from the top third. One study found that 44% of ChatGPT’s citations come from the first third of an article’s content. This means your most important, most citeable information needs to be front-loaded. If your key insight is buried in paragraph 12, ChatGPT might never get there. We restructured our own content production process based on this finding. Every piece we publish now puts the definitive answer, the original framework, or the key data point within the first 500 words.

Content length is nuanced. The raw word count correlation with AI citations is nearly zero. A 500-word post can get cited just as easily as a 3,000-word guide. But longer content (over 2,300 words) is 25 to 30% more likely to be cited than very short posts under 500 words, likely because longer content tends to cover more entities, include more specific data points, and provide more opportunities for AI systems to find citable passages.

The Platform Split Is Real

One thing that became clear over 12 months is that different AI platforms cite different types of sources. This matters because optimizing for one platform doesn’t guarantee visibility on another.

Google’s AI Overviews show the strongest preference for branded, commercial content. About 60% of AI Overview citations come from brand-controlled sources. Google leans toward sites it already trusts through its traditional ranking signals.

ChatGPT is more democratic. It pulls heavily from Wikipedia (about 8% of all citations), Reddit, Forbes, Medium, and PR Newswire. If your brand gets press coverage or publishes on platforms with high domain authority, ChatGPT is more likely to pick it up.

Perplexity leans hardest on Reddit and community-driven content. Nearly 47% of Perplexity citations come from Reddit. It also favors blogs, particularly product-focused and expert blogs. Perplexity is the platform where smaller, specialized brands can punch above their weight because it values depth and authenticity over raw domain authority.

Here’s the strategic implication: you can’t optimize for “AI search” as a monolith. You need a presence strategy that covers the citation patterns of each platform. That means your own website matters (Google AI Overviews), your LinkedIn presence matters (professional queries across all platforms), your Reddit presence matters (Perplexity, ChatGPT), and your press coverage matters (ChatGPT, Gemini).

This is why we tell clients that siloed marketing doesn’t work anymore. Your SEO agency can’t just optimize your website and call it done. Your social media manager can’t just post on Instagram and call it done. Everything feeds everything. Every platform is a potential citation source for every AI system.

What the Winners Are Doing Differently

After 12 months of tracking, a clear pattern emerged. The brands that are winning AI citations share five characteristics.

They publish consistently. Not sporadically. Not in bursts. Consistently. The AI systems reward publication cadence because it signals active expertise. A brand that publishes weekly builds entity authority faster than a brand that publishes a batch of content once a quarter and goes silent.

They show up across multiple platforms. The winning brands aren’t just on their website. They’re on LinkedIn. They’re in Reddit discussions. They’re getting quoted in industry press. They’re on podcasts. Every touchpoint builds the entity graph that AI systems use to determine citation-worthiness.

They create content worth citing. This sounds obvious, but most marketing content is not citeable. It’s generic. It’s a rehash of what everyone else published. It doesn’t contain original data, frameworks, calculations, or perspectives. The brands getting cited are the ones saying something nobody else is saying, with specificity and conviction.

They invest in structured data. Schema markup, author pages, clear entity relationships. These brands make it easy for AI systems to understand who they are, what they’re expert in, and why their content should be trusted. It’s the behind-the-scenes technical work that most businesses skip because it’s not glamorous. But the data is clear: sites with properly implemented schema are being cited at roughly 3x the rate of sites without it. You can have the best content in the world, but if AI systems can’t parse who wrote it and why it matters, they’ll cite someone else.

They build brand, not just content. This is the big one. The brands winning in AI search are the ones people already know and trust. Because AI systems are pulling from sources that other sources reference, brands with high name recognition and strong reputation signals get cited more frequently. It’s a compounding effect. Brand drives mentions. Mentions drive entity authority. Entity authority drives citations. Citations drive more brand awareness. The flywheel accelerates over time, and the brands that started building it early are nearly impossible to catch.

What This Means for Your Business

If you’ve read this far, you’re probably asking yourself one of two questions: “How do I start?” or “Am I already too late?”

You’re not too late. But you need to start now.

The AI citation economy is still in its early innings. Most businesses haven’t even heard the term “generative engine optimization,” let alone built a strategy around it. The window to establish authority is open, but it’s closing. As more brands figure this out and start competing for citations, the cost to break in goes up. The brands that build entity authority in 2026 will have a structural advantage that compounds every year.

Here’s the minimum viable strategy: publish original, data-backed content on your website at least twice a month. Post on LinkedIn at least three times a week with genuine expertise, not motivational fluff. Monitor and participate in Reddit communities relevant to your industry. Implement schema markup on every page of your website. Build your author profiles with verifiable credentials. Get your name mentioned in industry conversations, press, and communities.

This isn’t a marketing tactic. It’s a fundamental shift in how brands get discovered. The businesses that understand this will own the next decade of search. The businesses that don’t will wonder why the phone stopped ringing.

The Bottom Line

We’ve been building toward this for the last year. Everything we publish, every LinkedIn post, every piece of client content, every data-driven article on our site is designed to build the kind of entity authority that AI systems reward. Not because we predicted the future, but because we watched the data and followed where it pointed.

That’s what data-first marketing actually means. You don’t guess. You track, you analyze, and you go where the evidence says to go. The evidence says AI search is the future, and the brands building authority now are the ones that will own it.

A year from now, I’ll write this article again with another 12 months of data. The brands that started building today will be in it. Will yours?