How AI Models Really Choose Content in Search

In this episode, James Dooley speaks with Dan Petrovic about the evolution of AI SEO and how large language models are transforming search behaviour. They break down retrieval augmented generation, query fan out, selection rate optimisation and the importance of understanding model psychology. Dan explains how LLMs interpret and trim content, why traditional SEO foundations still underpin AI results, and how brands can test and strengthen their relevance within AI driven search environments. The discussion also covers probabilistic thinking, entropy, and practical ways to influence both grounded responses and long term model perception.