SEO Video Show · LiveAI Search & LLMsTechnical SEO

Generative Engine Optimization: State of AI Search | Andrew Yan

With Andrew Yan, Co-founder and CEO of Athena HQ, former Google Search product manager · Streamed live · 49 min

What this episode covers

Andrew Yan explains how generative engine optimization changes SEO measurement and proposes three AI search KPIs: brand mention rate, share of voice and citation rate. He covers how cited sources differ between AI models, what log files show about AI crawlers, and prompt planning as the new keyword research.

Key takeaways

  • ✦Track brand mention rate, share of voice and citation rate.
  • ✦Benchmark by industry, since each vertical behaves differently.
  • ✦Use log files to see which paths AI crawlers request.
  • ✦Find the most influential third-party sites for each AI model.

Key moments

  1. 19:57Three AI search KPIs
    “Three key KPIs, which is brand mention rate, share of voice, and then citation rate.”
  2. 22:44Benchmark by vertical
    “Each vertical needs to have its own benchmark.”
  3. 26:08What AI crawlers request
    “30 to 40% of the files that are accessed by AI coming from /blog.”
  4. 27:45Sources differ by AI model
    “ChatGPT cares more about Reddit, right? Versus on Gemini, we actually don't see Reddit ranked that highly.”

Quotes are from the episode transcript, lightly cleaned of auto-caption errors.

Questions answered in this episode

What three KPIs does Andrew Yan suggest for measuring AI search performance?
He suggests brand mention rate, share of voice and citation rate. Share of voice is relative to competitors, brand mention rate measures absolute performance, and citation rate measures whether your website informs the response.
Should SEOs optimize for agentic search now?
Not prematurely, because most consumers are not using agentic search today, so he advises taking it one step at a time.

More episodes on AI Search & LLMs