Yes. A landmark study by Similarweb (The Downstream Impact of AI Visibility) found that when a brand is recommended by ChatGPT, it is 2.5 times more likely to receive a website visit within the following 7 days than a brand that is not recommended [1].
The study measured consumer industries (finance, travel, and beauty), but the pattern matters for any marketer, including scientific B2B: most of that follow-on traffic does not arrive through an AI referral link. 55.9% of it arrives via branded search [1] — people typing the brand's name into a search engine after seeing it in an AI answer. Standard analytics record this as ordinary search, so the role AI played is hidden.
When someone is researching a purchase, they increasingly turn to conversational AI for initial ideas and validation. Similarweb's clickstream panel data shows what happens next: when an AI tool recommends a brand, visits and branded searches for that brand rise more than they do for brands that were not recommended [1].
The likely sequence is that the user reads the recommendation, leaves the chat, and later types the brand or product name into a search engine. The AI answer appears to influence the later visit, but traditional search gets the attribution [1]. As Rand Fishkin notes in SparkToro's analysis, this suggests AI answers have a real effect on buyer preferences in the industries studied — though he is careful to add that the effect is "small right now," is probably growing, and has not yet been tested for B2B or smaller brands [2].
Correlation vs. Causation: The study shows correlation, not proof that the AI recommendation caused the later visit [1][3].
Consumer vs. B2B Verticals: The data covers consumer verticals, not scientific or B2B markets. It is reasonable to expect a similar pattern among technical buyers — Fishkin's own audience research suggests higher-income, considered-purchase audiences lean slightly more on AI — but that is an informed expectation, not a measured finding [2].
For scientific B2B marketers, the practical takeaway holds either way: visitors who arrive after this kind of research tend to be more engaged. In the study, AI-influenced visitors viewed an average of 12.0 pages and spent 11.8 minutes per session, compared with just 6.5 pages and 5.6 minutes for other visitors [1]. They arrive further along in their research.
| Metric Evaluated | Finding & Data Points |
| Visit Probability | AI-recommended brands were 2.5x more likely to get a site visit within 7 days [1]. |
| How Traffic Arrives | 55.9% of downstream visits came via branded search, not direct AI referral links [1]. |
| Pages Per Session | 12.0 for AI-influenced visitors vs. 6.5 for others [1]. |
| Session Duration | 11.8 minutes for AI-influenced visitors vs. 5.6 minutes for others [1]. |
| Industries Studied | Finance, travel, beauty (US desktop); ChatGPT recommendations [1]. |
Note: Engagement figures are correlations. The study did not test scientific or B2B markets, and covered ChatGPT specifically rather than all AI tools [1][3].
Because most of the effect is hidden inside branded search. In the study, 55.9% of the follow-on traffic came from people searching the brand name after seeing it in ChatGPT [1]. Your analytics attribute those visits to normal search, so the AI's role is invisible. You are not seeing AI referrals because much of the impact is sitting quietly in your branded search volume.
AI recommendations can change from one query to the next — an earlier SparkToro analysis by Rand Fishkin found recommended brands often shifted across repeated versions of the same prompt [2][3]. Because you cannot lock in a single AI tool's output, a sound strategy focuses on broad Answer Engine Optimization (AEO) that improves your visibility across multiple models, rather than betting your entire budget on one platform.
Structure your content so AI tools can easily read, parse, and cite it. For scientific brands, that means making your data, application notes, and technical authority clear and structured. The goal is to be the foundational source an AI recommends, so that when a researcher later searches your name, they arrive at your site already informed and ready to engage.
https://www.similarweb.com/corp/the-downstream-impact-of-ai-visibility/https://sparktoro.com/blog/new-research-from-similarweb-how-ai-brand-mentions-influence-direct-visits-traditional-search-queries/https://www.searchenginejournal.com/ai-recommended-brands-saw-2-5x-more-site-visits-similarweb/580241/