Two sequential Generative Engine Optimization experiments point to an important adjustment for companies pursuing visibility in AI-generated answers: the pages that mention a brand may matter as much as, or more than, the brand's own website. The findings do not make owned content irrelevant. Instead, they suggest that early AI citations often depend heavily on third-party sources, making earned media, independent listicles, and external brand references a more central part of GEO strategy.
The experiments, published September 14, 2026, are detailed in Search Engine Land's analysis of the GEO experiments. They tracked brand mentions across multiple AI platforms and revisited the contribution of owned content. Their shared lesson is that businesses should not treat on-site publishing as the only route to appearing in AI answers.
What the GEO experiments found
The first experiment tracked brand mentions across several AI platforms. It found that a large share of mentions came from third-party pages rather than the company's own website. Owned pages contributed a smaller share of early citations.
The second experiment focused again on owned content, particularly listicles. It suggested that these pages work better as a foundational asset than as the primary engine of AI visibility. Third-party listicles continued to account for the majority of early mentions in the experiment.
| Area examined | Experiment 1 | Experiment 2 |
|---|---|---|
| Primary focus | Brand mentions across multiple AI platforms | The role of owned content, including listicles |
| Role of third-party pages | A large share of mentions came from third-party pages | Third-party listicles generated the majority of early mentions |
| Role of owned content | Contributed a smaller portion of early citations | Appeared to be a foundational asset rather than the primary driver |
These findings challenge a simplified version of GEO advice: publish a well-optimized page and expect AI systems to surface it. The research instead indicates that AI-answer visibility is shaped by the broader information environment around a brand. A company's claims on its own website can establish useful reference material, but independent sources may carry greater weight in early brand mentions.
That is a meaningful distinction for a business deciding where to spend limited marketing resources. It shifts attention from content production alone toward the combination of strong owned material, credible external coverage, and visibility into where AI systems are finding and citing the brand.
What businesses should measure
Traditional SEO reporting often centers on rankings, traffic, backlinks, and conversions. Those measures remain useful, but the experiments indicate that GEO reporting also needs to identify which sources associate a brand with a topic in AI-generated answers.
A practical measurement approach should distinguish between:
- Mentions or citations from owned pages and those from third-party pages.
- The external publications, listicles, and other pages that repeatedly surface alongside the brand.
- The topics and questions for which a brand appears in AI answers.
- Changes in early AI mentions after new external coverage or owned content is published.
The available research does not establish a universal formula for all AI platforms, industries, or queries. It does, however, provide a reason to stop using website content as the sole proxy for AI visibility. A report that only measures on-site content output or conventional search performance can miss the external signals that the experiments found to be influential.
How content and PR strategy should change
The implication is not to abandon owned content. Businesses still need accurate pages that clearly explain their products, services, expertise, and differentiators. Those pages can provide the foundation for both readers and external publishers. But the research suggests that publishing another company-controlled listicle, without a plan to earn independent mentions, may have limited impact on early AI citations.
A more balanced strategy can connect content and PR work. For example, a company can use its owned material to supply clear facts and useful expertise, then pursue legitimate opportunities for relevant third-party coverage. The goal is not to manufacture citations or treat every listicle as equally valuable. It is to ensure that independent, topic-relevant sources can accurately describe the brand and its place in a category.
For budget planning, this creates a case for reviewing whether all available resources are flowing into blog production and technical on-site optimization. The experiments suggest that earned media and external references deserve a defined place in the mix, alongside the owned pages that support them. The right allocation will vary by business, because the research does not prescribe a fixed spending ratio.
For companies that want to understand whether external sources, not just their own pages, are shaping their AI presence, Scalevise can help map the questions, citations, and brand signals that matter through its AI Visibility and GEO Checker. That visibility can help teams focus content and PR effort on gaps that affect how their brand is represented in AI answers, rather than relying on publishing volume alone. Start an AI Visibility scan.
A more useful way to evaluate GEO tooling
The experiments also raise the bar for GEO tools and internal reporting. A useful workflow should not only flag whether a brand appears in an AI response. It should help identify the third-party sources associated with that appearance and show where owned pages play a supporting role.
That creates a clearer feedback loop. Marketing teams can compare the external sources appearing for priority topics, assess whether the brand is represented accurately, and decide whether the next investment belongs in an owned resource, a PR initiative, or both. The central lesson is that AI visibility is not confined to a company's domain.
Frequently Asked Questions
What did the GEO experiments find about third-party sources?
The experiments found that a large share of early brand mentions in AI platforms came from third-party pages. In the second experiment, third-party listicles generated the majority of early mentions.
Does this mean businesses should stop investing in owned content?
No. The research suggests owned content is a foundational asset, not necessarily the primary driver of early AI visibility. Accurate owned pages can support a broader strategy that also earns relevant third-party mentions.
What should a GEO measurement program track?
It should distinguish owned from third-party citations, identify external sources associated with the brand, and monitor the topics and questions where the brand appears in AI-generated answers.
Do the experiments prove the same outcome for every AI platform and industry?
No. The research reports findings from the experiments across multiple AI platforms, but it does not establish a universal formula for every platform, industry, or query.
Conclusion
The two GEO experiments make a clear strategic case for looking beyond a company's own website. Owned content remains important, but early AI mentions in these tests were more strongly associated with third-party sources. Businesses evaluating their AI visibility should therefore measure external brand signals, coordinate content with credible PR activity, and avoid treating on-site publishing as a complete GEO strategy.
