G2’s AI citation study is a land grab disguised as research, but it reveals something real: the infrastructure of B2B buyer influence is being rebuilt around which sources AI models trust at the bottom of the funnel.
G2 recently published a study analyzing roughly 35,000 ChatGPT citation URLs to answer a straightforward question: do review platforms get cited more as buyer intent increases? The answer, based on their data, is yes. Review platform citations nearly doubled from the discovery stage (7.4%) to the evaluation stage (13.2%). And within those review citations, the G2 portfolio (including its newly acquired Capterra, Software Advice, and GetApp) commands 84% share.
The study is interesting. But not for the reason G2 wants it to be.
The real game being played
Set aside the specific numbers for a moment. What G2 actually did here is publish a framework for how influence works inside AI-generated recommendations. They’ve mapped the buyer journey onto citation patterns and staked a claim: when it matters most, AI turns to us.
This is a land grab disguised as research.
And it’s a smart one. Because in traditional search, the proxy for influence was rankings. In AI search, the proxy is citations. G2 is essentially arguing that citations are the new rankings, and that they own the bottom of the funnel.
What the data actually shows (and doesn’t)
The methodology is worth examining. The study covers US-only ChatGPT citations from December 2025, sourced through Profound. That’s a real dataset, but it’s a narrow one. It doesn’t cover Perplexity, Gemini, or Claude. It doesn’t account for how citation patterns vary across AI models. And the intent classification (discovery, exploration, evaluation, focused evaluation) was applied to the prompts after the fact, which means the buckets themselves are a design choice, not an objective measurement.
None of that makes the finding wrong. Review sites probably do show up more in evaluation-stage queries. That passes the common sense test. When someone asks ChatGPT for the best project management tool for a 50-person engineering team, the model is going to lean on structured comparison data. That’s what review platforms have.
But the jump from “review platforms get cited 13% of the time during evaluation” to “review sites are most effective at the bottom of the funnel” is doing a lot of work. Vendors still dominate citations at every stage, accounting for 67-74% of all cited domains. The real story might be that vendor sites are the citation baseline and everything else fights over the remaining 30%.
The citation economy is forming right now
Here’s what’s actually worth paying attention to: a new influence layer is consolidating, and most B2B companies haven’t noticed.
In the old model, a buyer searched Google, saw organic results, maybe clicked a G2 comparison page, and eventually landed on a vendor site. The vendor could track that journey. They could optimize for it. They could buy ads against it.
In the AI model, the buyer asks a question and gets an answer. The citations underneath that answer determine which vendors get mentioned and which evidence the model treats as credible. The buyer might never visit G2 at all. But G2’s data shaped the recommendation.
This means G2’s real product is shifting. It’s becoming less about driving referral traffic to vendors and more about being the training data and citation source that AI models trust. The acquisition of Capterra, Software Advice, and GetApp makes more sense through this lens. It’s not just about consolidating review market share for human buyers. It’s about ensuring that when any AI model needs structured, comparative software data, the G2 ecosystem is the default source.
What this means for vendors
The practical implication is uncomfortable. If AI citations are becoming a bottom-of-funnel influence channel, then your G2 profile isn’t just a nice-to-have for prospects who browse review sites. It’s a data input that shapes whether AI recommends you at all.
That changes the ROI calculation for review generation programs. It changes how you think about the content on your G2 profile. And it raises a question that most B2B marketing teams haven’t started asking: what does your product look like to an AI model that’s deciding what to recommend?
G2 wants the answer to be: invest more in G2. That’s the obvious play. But the deeper point stands regardless of who benefits. The infrastructure of buyer influence is being rebuilt around AI citation patterns. The companies paying attention to that shift now will have a structural advantage. The ones still optimizing purely for Google rankings are building on a foundation that’s already starting to shift.