iPullRank Experiment Reveals LLMs' Use of Private Data Sources, Challenging Universal "AI Share of Model" Metric
📊 In a recent experiment by iPullRank, led by Garrett Sussman, fabricated brands added to a Gmail inbox saw a 46-point lift in AI search visibility. This shows LLMs (Large Language Models) now access private data sources like emails, challenging the accuracy of the universal "share of model" score touted by the GEO industry. Consequently, email marketing and loyalty programs are emerging as key areas for LLM discovery, beyond public web content.
Share