Traditional SEO is no longer enough. Companies must understand how AI search platforms like ChatGPT, Copilot, Gemini, and Perplexity surface and rank content. Rankscale founder Mathias Ptacek explains why AI visibility requires new strategies.
As AI-driven search platforms reshape how audiences discover content, brands and publishers face new challenges in staying visible. Mathias Ptacek, founder of the Austrian startup Rankscale, has focused on analyzing how often and where brands, products, and content appear in answers from ChatGPT, Copilot, Gemini, and other AI systems. He said that classic SEO tactics alone no longer guarantee visibility in these environments.
Ptacek began exploring AI search visibility in mid-2024 while leading a software company. His initial goal was to improve his firm's presence on Perplexity, which led to building a tool that analyzed website performance in AI search results. This side project evolved into Rankscale, a platform that now tracks and measures brand mentions and positioning across major AI search engines.
Rankscale operates with a small team and has grown without venture capital, relying on strategic investors and business angels. Ptacek said the company is focused on defining which market segments to serve, including AI commerce, publishing, and brand visibility. He noted that while the sector is experiencing significant hype, the company’s agility allows it to move quickly and adapt to new developments, unlike larger SEO platforms that have been slower to react.
According to Ptacek, understanding how AI systems select and cite sources is difficult because these platforms function as black boxes. Rankscale approaches this by sending large sets of prompts to AI systems and statistically analyzing the responses. The company examines which sources are cited, how brands are described, and the sentiment of those mentions. Ptacek observed that the importance of specific sources varies widely between models; for example, a source prominent in ChatGPT may be nearly absent in Gemini.
He emphasized that companies must monitor not only their own content but also identify authoritative third-party sources in their field. By targeting these sources through PR, partnerships, or contributions, brands can increase their chances of being cited by AI systems. Ptacek said that structured information pages, such as factsheets or landing pages, help AI models better understand and surface brand information.
Prompt sets are central to Rankscale’s methodology. Ptacek explained that measuring visibility with only one type of query gives a limited view. Instead, prompt sets should reflect different search intents-informational, transactional, and commercial-to provide a realistic picture of brand presence. Rankscale’s tools help companies build and refine these prompt sets by analyzing topics, markets, and languages to identify the most relevant queries.
Ptacek reported that the sources cited by AI systems differ significantly. For instance, Copilot often references platforms like Semrush and Ahrefs, while ChatGPT relies more on Reddit, Ars Technica, or TechRadar. He also noted that Copilot tends to cite fewer sources overall but may feature certain platforms more prominently.
To optimize for AI search, Rankscale has created dedicated facts pages designed primarily for LLM crawlers rather than human visitors. These pages are heavily used by AI systems such as Grok and Google’s AI products. Ptacek advised that brands should create content that directly answers specific questions, ideally matching the query in the URL and providing clear, structured information.
He also highlighted that some early SEO strategies-like building pages that answer precise questions-still work well in the AI era. However, he cautioned that manipulative tactics, such as always placing one’s own brand at the top of listicles, may be detected and lose effectiveness over time.
Rankscale’s reporting goes beyond simple mention tracking. The platform calculates a visibility score that combines detection rate and position within AI-generated answers. Ptacek said this approach provides a more nuanced view of brand prominence than tools that only count mentions. The system also analyzes which sources AI systems rely on and tracks so-called “fan-outs”-additional internal queries generated by AI models in response to a user’s question. These insights inform actionable recommendations for PR, content optimization, and earned media strategies.
For small and medium-sized businesses, Ptacek recommended focusing first on offsite visibility by targeting key authoritative sources, then optimizing on-site content for AI comprehension. He said that clear, well-structured pages that answer specific questions remain highly effective for AI search visibility.
Ptacek acknowledged ongoing debates about content licensing and attribution in AI search. He noted that while being cited as a source can benefit brands, there is currently no guaranteed right to a link or attribution when AI systems use content. He also observed that changes to authoritative sources can quickly influence AI-generated answers, making the environment highly dynamic and sometimes vulnerable to manipulation.
These shifts in AI search are prompting publishers and brands to rethink their strategies. As seen in other recent developments, such as major tech companies integrating agentic AI into media workflows, the relationship between content creators and AI platforms is rapidly evolving.
