SEO Video Show · LiveAI Search & LLMsLocal SEO

AI Search Optimization: How to Test ChatGPT and Gemini | Sanjay Singh

With Sanjay Singh, Founder of Radar Kit AI and Visibly AI · Streamed live · 50 min

What this episode covers

Sanjay "Jay" Singh explains how to test AI visibility properly: run each prompt many times and analyze the answers as a pool, and pull the real fan-out queries from ChatGPT's network tab instead of guessing them. He also shares a LinkedIn test that got him cited.

Key takeaways

  • ✦Test each AI model separately, since their data partners and cited sources differ.
  • ✦Inspect network traffic in AI interfaces to see the searches behind answers.
  • ✦Analyze hundreds of answers before drawing conclusions.
  • ✦LinkedIn articles can surface quickly in AI answers for B2B topics.

Key moments

  1. 18:32Find fan-outs in the network tab
    “Just go to the network tab and then you will see”
  2. 20:03Analyze hundreds of answers
    “You need at least hundreds of sources, hundreds of answers analyzed.”
  3. 24:41LinkedIn articles in AI answers
    “I started writing articles on my LinkedIn.”
  4. 32:04Reviews as fan-outs
    “Reviews are the most used fan-outs.”
  5. 33:58Extract, don't generate
    “Query fan-out cannot be generated. It should be extracted.”

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

Questions answered in this episode

How can I see which sources ChatGPT uses to answer a query?
Jay says you can open the network tab in the ChatGPT interface and look in the conversations folder to see query fan-outs, reasoning, and web searches performed for your prompt.
Do LLM answers stay the same when you ask the same question repeatedly?
No, Jay says the answers are probabilistic, so the brands mentioned stay similar but their order shuffles, which is why he recommends running a prompt many times to build a baseline.
Can I track my business in AI map results for local SEO?
Jay says his Local Radar feature tracks map listings shown in AI answers and compares appearance rates against competitors.

More episodes on AI Search & LLMs