By Tim Peterson * October 30, 2025 *
Ivy Liu.
Artificial intelligence platforms such as OpenAI’s ChatGPT and Anthropic’s Claude currently contribute a modest single-digit percentage to Forbes’ monthly referral traffic. However, their impact extends beyond mere pageviews; these AI-driven sources have played a crucial role in mitigating the significant 40% decline in search referral traffic Forbes experienced this year.
Forbes leverages detailed analytics on AI referral traffic, including the specific prompts that lead users to cite Forbes content. This data enables the creation of distinct audience segments based on visitors arriving via AI platforms, offering valuable insights into user behavior and interests.
“Tools from providers like Semrush and Similarweb now offer insights into which prompts are driving traffic to your content and quantify that traffic,” explained Forbes Chief Innovation Officer Nina Gould during the Digiday Publishing Summit Europe in Lisbon on October 29. She noted that such data is not universally accessible across all AI platforms. For instance, Google’s AI Mode remains opaque, as referral data from its ChatGPT-style search experience is indistinguishable from traditional search metrics, making it a “black box.”
Excluding Google AI Mode, Forbes utilizes data from Semrush and Similarweb to build audience cohorts, categorizing visitors based on their AI-driven search behaviors.
“We develop cohorts to understand the nature of searches and the interests of our visitors. Importantly, we do not track or target individuals upon their return, so there are no concerns regarding personally identifiable information,” Gould emphasized. This aggregated data informs editorial teams about which Forbes content resonates most with AI platform users-content that prompts users to click through from ChatGPT, Perplexity, and similar services. Gould identified Forbes’ “core” content as including lists, wealth reporting, entrepreneurship, and investment topics.
Additionally, Forbes has identified distinct behavioral patterns among AI-driven visitors.
“These users often expect highly personalized experiences and concise information. They seek comprehensive insights delivered quickly and efficiently in a single location,” Gould observed.
In response, Forbes has enhanced its website to better serve this audience.
For example, the “More For You” article recommendation module, previously based solely on contextual relevance, now incorporates user-based personalization. This change has resulted in a remarkable 300% increase in clickthrough rates, according to Gould.
This personalization stems from a broader redesign of Forbes’ homepage, which was restructured to be modular and customizable, allowing elements from the homepage to be integrated into article pages. Following this update, Forbes experienced a 45% uplift in clickthrough rates.
As AI platforms continue to evolve, publishers like Forbes must explore innovative strategies to adapt their digital properties for the AI era.
Ensuring accurate attribution of their reporting by AI platforms is critical. A recent study by the BBC, European Broadcasting Union, and British Council revealed that over one-third of UK adults hold publishers accountable for content disseminated via AI, even when errors occur.
To address this, publishers could benefit from implementing enhanced HTML schema markup that clearly signals article content to AI systems. This approach would be analogous to existing schemas used to optimize content for platforms like Google and Meta, providing a standardized method for AI platforms to identify and cite accurate information.
Unfortunately, no universal standard currently exists, nor is one actively being developed. Given the rapid acceleration of AI adoption, establishing such standards is urgent.
“Our business models are evolving at an unprecedented pace, and the tactics required to maintain visibility and competitiveness in this new landscape are constantly shifting,” Gould concluded. “We are learning and adapting in real time.”
More in Media.
Dispelling AI Content Myths: Separating Fact from Fiction and Noise
The term “AI Slop” has become a catchphrase for low-quality AI-generated content due to its simplicity and memorability. However, this label oversimplifies the complex and evolving landscape of AI content quality.
Insights from the October 2025 Digiday Publishing Summit Europe revealed publishers’ concerns about the long-term recovery of web traffic during a private session.
Exploring how agencies, platforms, and publishers are integrating AI agents into their workflows.
