Join the authors of Beyond Search: LLM Adoption and the Concentration of Web Traffic for a 45-minute DCN Research Briefing and discussion.

Samira Gholami (Stanford University), Cristiana Firullo (Cornell University), Cristobal Cheyre (Cornell University), and Alessandro Acquisti (MIT Sloan School of Management) will present new empirical findings on how large language models are affecting search behavior, web exploration, and the distribution of traffic across external sites. The research was recently presented at the Digital Competition Conference 2026.

What you’ll learn:

Does AI reduce open web traffic?
Early evidence suggests LLM adoption is associated with a broader set of websites visited, challenging assumptions about traffic contraction.

How are audiences actually using LLMs?
LLMs are most often used alongside traditional search, not as a replacement. Mixed sessions that combine search and LLM use are longer and more exploratory.

How is traffic being redistributed?
While Google, Bing, and LLM platforms all send users externally, LLM-driven traffic appears more widely distributed rather than concentrated among a few dominant sites, with implications for discovery and audience strategy.