Data Breakdown · AI Search Citations and Sources
Why third-party pages drive 85% of brand mentions in AI search
An analysis of AirOps' study of 21,311 brand mentions across ChatGPT, Claude and Perplexity, and what it means for how software companies earn AI visibility.
- Original research
- AirOps
- Published
- Last updated
- , sources checked 5 November 2026
- Analysis by
- Asad Shahzad · 4 min read
This is AEO Operators' analysis of research published by AirOps. It is not original research by AEO Operators.
Executive summary
AirOps found that 85% of brand mentions in AI answers to discovery-stage software questions came from third-party pages, and that nearly 90% of those came from listicles, comparisons or reviews.
For a software company, this means your own website sets the facts, but independent lists and reviews may influence whether buyers encounter your product during early evaluation; this study does not establish that they determine shortlist inclusion.
Research background
AirOps published Third-Party Sources Drive 85% of Brand Discovery on 17 October 2025.
It analysed 21,311 brand mentions in answers to more than 500 commercial-intent queries in software verticals including MarTech, Sales, HR and Finance, and Productivity and Collaboration.
Each query was tested on GPT-5, Claude Sonnet 4.5 and Perplexity Sonar. The queries focused on the discovery stage, when a buyer is trying to find out which brands exist in a category.
How the research was done
AirOps ran each query on three models and classified the source of every brand mention as the brand's own domain, a third-party domain, or uncited.
It then classified the type of content behind each mention and checked which brands appeared on more than one platform.
- Report
- Third-Party Sources Drive 85% of Brand Discovery
- Research period
- Published 17 October 2025
- AI platforms
- ChatGPT, Claude, Perplexity
- Original author
- AirOps
Key research statistics
21,311
brand mentions analysed across three AI platforms.
More than 500 commercial-intent queries. Source: AirOps
~90%
of third-party mentions came from listicles, comparisons or reviews.
Reported as nearly 90%. Source: AirOps
The numbers in charts
Each chart uses the values published in the original report.
Where brand mentions in AI answers came from
Share of 21,311 brand mentions by source
Bar chart. AirOps reports 85% of mentions from third-party domains, 13.2% from owned domains and about 2% uncited. Rounded figures do not sum exactly to 100%.
Show the data table
| Source | Share of mentions |
|---|---|
| Third-party domains | 85% |
| Brand's own domain | 13.2% |
| Uncited | ~2% |
Method and limits: AirOps reports rounded shares of 85%, 13.2% and about 2%, which total 100.2%. These are not exact components of a 100% whole; no residual has been invented.
Research period: Published 17 October 2025. Sample: 21,311 brand mentions, 500+ queries.
Values marked with ~ are approximate as published or read from the original chart.
Data source: AirOps, Third-Party Sources Drive 85% of Brand Discovery, 17 October 2025. Visualization recreated by AEO Operators. Values are approximate reconstructions from the published chart. View the original report
Bar chart. AirOps reports 85% of mentions from third-party domains, 13.2% from owned domains and about 2% uncited. Rounded figures do not sum exactly to 100%.
Owned pages behind first-party mentions
Share of first-party mention content by page type
Horizontal bar chart. Product pages made up 19.3% and homepages 7.1% of first-party mention content.
Show the data table
| Page type | Share of first-party content |
|---|---|
| Product pages | 19.3% |
| Homepages | 7.1% |
Method and limits: AirOps reports only these two page types on the report page, so other page types are not shown.
Research period: Published 17 October 2025. Sample: First-party mentions within 21,311 brand mentions.
Data source: AirOps, Third-Party Sources Drive 85% of Brand Discovery, 17 October 2025. Visualization recreated by AEO Operators. View the original report
Horizontal bar chart. Product pages made up 19.3% and homepages 7.1% of first-party mention content.
What the research actually reveals
The study is unusually relevant to us because its queries are about software categories, which is exactly the situation our clients face.
AirOps' conclusion is that owned and earned content work together, and we agree, but the balance is striking: 13.2% of mentions came from the brand's own domain.
The finding that, among the third-party listicle, comparison and review instances, 80% of brands often appeared among the first three companies discussed in the cited content is, in our view, the most practical one.
It suggests that being listed is not enough, and that position within a list matters too.
The platform differences matter as well. About 68% of brands appeared on only one platform, and GPT-5 relied least on first-party content, at 4% to 11% depending on category.
A company that tracks only one AI platform could reach very wrong conclusions about its visibility.
Implications for Series A to Series C SaaS companies
Founders should treat inclusion in credible category lists, comparison sites and review platforms as a core visibility task, not a side project for public relations.
Growth and product marketing leaders should keep product pages and the homepage precise, because together they made up more than a quarter of first-party visibility.
Heads of Growth should measure visibility on at least three AI platforms, since most brands in the study appeared on only one.
Actionable recommendations
- List the ten most-cited comparison articles and review pages for your category and check whether you appear, where you appear and whether the description is correct.
- Prioritise getting into the first three positions on lists you already appear in by giving authors fresh, verifiable proof points.
- Make your product pages state clearly what the product does, who it is for and how it differs, because product pages accounted for 19.3% and homepages for 7.1% of first-party mention content in the study, or 26.4% combined rather than a majority.
- Track ChatGPT, Claude and Perplexity separately and report the overlap, not just a combined score.
- Keep messaging consistent across your site, review profiles and partner pages so that every source repeats the same facts.
Research limitations
AirOps has a commercial interest in content-led visibility. The study covers discovery-stage commercial queries in a limited set of software verticals, so it may not describe late-stage or brand-name queries.
It does not state the exact collection dates or the geographic setting of the queries. The model versions tested will change over time, and the results describe correlation rather than cause.
Original sources and further reading
- Third-Party Sources Drive 85% of Brand Discovery
AirOps, AirOps, 17 October 2025. Section: Whole report. Vendor research.
Primary source for every figure in this article.
- The 2026 State of AI Search
AirOps, AirOps, 2 December 2025. Section: Off-site presence section. Vendor research.
Repeats the 85% and 6.5× findings in a wider summary.
Want to understand how your SaaS brand appears across AI search platforms?
Get an AI Visibility Assessment from AEO Operators, and we will apply what this research shows to your own company and competitors.
Get an AI Visibility Assessment