AI Search Requires a New SEO Mental Model: Insights from Tom Capper

Tom Capper’s AMA reveals that AI search is not a replacement for lexical search but a shift in how visibility is measured. Grounding searches, where an LLM pulls live data from the web, become the new ranking signal, meaning brands must ensure their content is accessible to grounding bots rather than treating all bots as blockers. Prompt research should move from keyword volume to topic‑level fan‑out analysis, tracking which themes AI systems cite and how those citations evolve over time. The LLMs.txt file, once dismissed as snake oil, is largely ignored by major AI bots, so its implementation offers little benefit for Google‑centric SEO. Success metrics shift from click‑through rates to brand visibility, recommendation, and accurate association, and off‑site digital‑PR content now has a larger impact than simply rewriting on‑page material. Training bots, which feed the model’s base knowledge, update slowly and blocking them rarely changes current AI responses, whereas grounding bots must remain open to allow real‑time retrieval. Tools like the Moz AI Visibility platform can map brand citations across grounding queries, helping teams monitor visibility at the topic level.