TL;DR:
Review-platform citation share increases meaningfully with stronger purchase intent, based on an analysis of ~35,000 ChatGPT citation URLs captured in Profound. Within these citations, G2 and its recently acquired brands — Capterra, Software Advice, and GetApp — collectively command 84% share of citations in the review-platform category. This confirms that as buyers move deeper into the funnel, AI models increasingly rely on the verified peer-generated sources found across the G2 portfolio to inform their recommendations.
Why this matters
The buyer journey no longer starts with a demo request; it starts with a prompt. Because citations are one of the few observable evidence layers behind an AI recommendation, they represent where a brand must show up to be considered.
AI answers mention software vendors by design. The more interesting question is which third-party sources those answers cite. If review platforms are cited more as intent strengthens, that supports the idea that verified reviews and buyer proof become more influential deeper in the funnel.
Hypothesis
As intent strength increases, the probability that a review site (such as G2 or TrustRadius) is cited increases more than for vendor sites or other third parties.
Methodology
Data and scope:
- Source: ~35,000 Profound citations
- Coverage: US only, ChatGPT only
- Time window: December 2025
1) Assign intent to each prompt
Each prompt was classified into one mutually exclusive user journey step (for example: discovery, exploration, evaluation, focused evaluation, or other), then normalized into a prompt key and mapped to a single intent.
|
Journey Step |
Description |
Intent strength |
# prompts |
% |
|
Discovery |
“What is …?” style, learning what a category or concept is |
Low |
376 |
11.8% |
|
Exploration |
“Which …” questions about integrations, capabilities, tradeoffs, usually seeking 1–few fits |
Medium |
1,179 |
37.1% |
|
Evaluation |
Generic “best/top” lists without strong constraints |
High |
1,367 |
43.0% |
|
Focused_evaluation |
“Best/top” with explicit constraints (segment, use case, environment, performance, etc.) |
Very high |
255 |
8.0% |
2) Convert citations into analyzable units at the domain level
For each citation URL, I extracted the root domain and classified it into a site type:
- review_platform (G2, Capterra, GetApp, Software Advice, TrustRadius, Gartner as review context)
- ugc_platform (Reddit, Wikipedia, Quora, YouTube, LinkedIn)
- publisher (media, blogs, editorial sites)
- vendor_or_other (vendors plus everything not mapped explicitly)
3) De-duplicate domains within each run
- To avoid a single response over-weighting a domain via multiple URLs, I de-duplicated to one record per run_id × intent × domain.
4) Compute shares
- Primary view: For each intent stage, compute the share of unique cited domains by site type.
- Secondary view: Within the review_platform subset, compute share by review domain (G2 vs Capterra vs others), overall, and by intent.
Results
A) Citation mix by site type
Vendors dominate citations in every stage (about 67% to 74%). That makes sense because the prompts we looked at asked for vendors.
|
Intent |
Publisher |
review_platform |
ugc_platform |
vendor_or_other |
|---|---|---|---|---|
|
Discovery |
4.5% |
7.4% |
17.8% |
70.4% |
|
Exploration |
2.4% |
7.7% |
18.2% |
71.8% |
|
Evaluation |
4.9% |
13.2% |
15.1% |
66.7% |
|
Focused_evaluation |
4.7% |
8.4% |
17.2% |
69.6% |
|
Overall |
4.0% |
8.6% |
17.1% |
70.4% |
However, the closer we get to a purchase in the buyer journey, the more prevalent and influential platforms like G2 become, thanks to a significant share of authentic peer reviews and UGC. Review-platform citation share rises to 13.2%, roughly 1.8× discovery’s 7.4%. Meanwhile, vendors see lower influence on those decisions at this same stage.
Review platforms are cited more than publishers (8.6% vs 4.0%).
B) Review-platform share split by brand
Within review-platform citations, G2 is the most frequently cited domain by a wide margin in this sample and timeframe.
Since G2’s acquisition of Capterra, Software Advice, and GetApp, it has established a dominant presence in AI-driven discovery. Here is the combined overall share across major review domains:
- G2 ecosystem (g2.com + acquired brands): 84%
- gartner.com: 13%
- sourceforge.net: 2%
- trustradius.com: 1%
- trustpilot.com: ~0%
This indicates that AI citations from review sites are a strong bottom-of-funnel lever for software vendors, particularly when they come from a major player like G2.
Win with trust
The data confirms that review sites are most effective at the bottom of the funnel. When software buyers reach the evaluation stage of their journey, the recommendations they see from AI increasingly stem from trusted peer proof sourced from platforms like G2.
With G2’s now significant influence on LLM citations and AI search visibility, there is a strong case for vendors to maintain a steady presence across the G2 ecosystem. This ensures that when AI looks for the final layer of evidence to recommend a product, it finds a wealth of authentic customer voices.
![Do Backlinks Still Matter in AI Search? Insights from 1,000 Domains [Study]](https://mgrowtech.com/wp-content/uploads/2025/10/do-backlinks-still-matter-in_-ai-search-study-sm-120x86.png)
