The rise of AI-powered search systems is transforming how brands gain visibility online. A 16-month experiment testing the performance of AI-discovered content from a newly created fictional brand reveals critical insights into the inner workings of AI search. Exploring how different AI engines index, cite, and rank this freshly launched brand provides valuable lessons for marketers aiming to shape their narratives in the era of large language models.

Experiment Overview and Methodology

The research team launched a fictional brand, creating a new website specifically for the experiment alongside 11 established domains aged over a year and with existing rankings. Across these properties, seven content formats were published ranging from deep guides and review articles to listicles and clickbait-style posts. The content output began in March 2026 and was tracked for AI mentions across five major AI systems: ChatGPT, Google AI Mode, Google AI Overviews, Perplexity, and Gemini. Over 15,800 AI-generated answers were analyzed from 825 unique prompts to assess brand visibility, citation frequency, and ranking positions within AI responses.

Key Findings: Branded vs Non-Branded Queries

One of the most significant discoveries was that 96% of AI visibility originated from branded searches. This indicates new domains face difficulty competing on broader, generic queries. However, on uniquely brand-related queries where the fictional brand was the only authoritative source—for instance, about product-specific claims—the visibility skyrocketed, outperforming established sites by as much as 32 times. The conclusion mirrors traditional SEO principles: a new brand must first build trust and recognition to challenge existing competitors on generic topics. AI systems appear to rely on authoritative, well-established sources for general questions but readily cite brand-owned content on unique brand claims.

Importance of Defining Brand Narrative in AI Search

Content that clearly defined the brand’s offerings and differentiators, such as ‘About Us’ and ‘Complete Guide’ pages, became the most cited sources. These pages cumulatively accounted for thousands of AI citations, demonstrating that even without domain authority, carefully positioned brand content can shape the AI narrative rapidly. Businesses should ensure their core pages comprehensively answer key questions about who they are and what they offer to influence AI search results for branded queries.

Variations in AI Engine Behavior

Distinct differences emerged between the AI systems in how quickly they indexed new content, cited sources, and maintained visibility.

Google AI Mode: The Most Stable Citation Source

Google AI Mode was the most reliable for maintaining top position citations on branded queries, consistently placing the fictional brand first around 90% of the time. This stability contrasts with other engines and provides a benchmark for dependable AI visibility.

Google AI Overviews: Fluctuating but Accurate

Google’s AI Overviews surfaced the brand frequently but with uneven citation consistency, showing considerable fluctuations in links and visibility over time. Despite variable presence, when citations appeared, the brand was described accurately, highlighting a potential tradeoff between consistency and fresh content updates.

Perplexity: Rapid Content Uptake but Fragmented Citation

Perplexity rapidly cited new content within 1 to 3 days of indexing, outperforming others in speed. However, Perplexity often spread citations across multiple supporting domains rather than centralizing visibility on the main brand site. This shows how freshness can boost visibility but may dilute brand authority if content is scattered.

ChatGPT: Building Visibility Over Time

ChatGPT showed slower initial recognition but steadily increased brand mentions and citations over the month. Uniquely, it displayed strong presence in review and comparison article formats, indicating that detailed content in these styles may help shape responses in conversational AI settings.

Gemini: Limited Recognition and Citation

Gemini underperformed compared to other engines, struggling initially to identify the brand’s niche and frequently omitting citations even in uniquely branded queries. Prompt structure influenced results somewhat, with comparative queries yielding better recognition, but overall Gemini showed the least consistent brand visibility.

Content Formats and Volume Matter

The experiment confirmed that content depth and structure significantly influence AI citations. Deep guides of 5,000 to 6,000 words generated around 900 AI citations per page, while review and comparison articles also performed strongly. In contrast, how-to articles and listicles saw minimal citation impact. Interestingly, publishing a large volume of shorter, thinner content (approximately 500–750 words) collectively generated the highest citation counts domain-wide, surpassing individual long-form content. This illustrates that while quality content drives AI citations, quantity can compensate by increasing the likelihood of retrieval and citation, especially in freshness-prioritizing engines like Perplexity.

Topical Clustering Alone Does Not Guarantee AI Visibility

Contrary to expectations rooted in traditional SEO, organizing content into a topical silo of a hub page linked to supporting articles did not generate any AI citations despite full indexing and internal linking. This result underscores that topical clustering, while useful, is insufficient on its own for AI visibility. AI systems prioritize content retrieval based on clear reasons for citation rather than purely semantic or structural signals.

Implications on AI’s Approach to Truth and Verification

AI visibility favored content that was consistent, repeated, and clearly presented over strict factual verification. The fictional brand received positive recommendations from some AI answers despite its artificial origin, reflecting that when negative information is absent, AI may fill gaps with neutral or favorable interpretations. This demonstrates that AI can be highly sensitive to manipulation and highlights the necessity for companies to proactively manage the information environment AI systems rely on.

“AI systems follow identifiable and testable signals, but the same factors that help legitimate brands become visible can enable simulation of credibility.” – Industry AI Analyst

Conclusion and Strategic Takeaways

The experiment reveals that AI search is neither random nor entirely objective. Brand visibility depends on strategic inputs like branded context, unique claims, diverse content formats, and distributed presence across multiple domains. Marketers must actively craft comprehensive, authoritative content that clearly presents brand narratives to shape AI search outcomes. Equally important is monitoring how different AI engines behave and tailor strategies accordingly. As AI search evolves, ongoing research and adaptation will be essential for maintaining an authoritative presence in AI-driven information ecosystems.

Additional Resources and Expert Recommendations

Marketers seeking to optimize for AI search should focus on creating detailed, unique brand-centric content promptly after launch. Establishing a variety of authoritative formats like deep guides and reviews accelerates AI recognition. Monitoring AI engine behavior using analytics tools can identify optimal content types and prompt structures for visibility. Additionally, understanding that volume of content influences retrieval in engines prioritizing freshness suggests a balance of quality and quantity can be effective.

Tools and platforms specialized in AI result tracking enhance the ability to measure brand presence and citation stability, enabling ongoing content optimization in response to AI algorithm changes.