Sat through 6 "AI search optimization" pitches this month. They all sell a "visibility score." Nobody can explain how it's calculated. What's actually the real methodology?
Hopefully, this saves someone else some hours.
Ran a small RFP last month for AI search visibility services. Mid-market B2B, trying to figure out whether we show up in ChatGPT / Perplexity / Gemini when someone asks about our category. Six agencies pitched.
Every single one of them had a proprietary "AI visibility score." Every single one. When I asked how it's actually calculated:
- two said they couldn't share without an NDA
- three basically said "we ask the LLMs a bunch of prompts and count how often you appear"
- one had a slightly more technical answer but couldn't tell me how many prompts they run, how often they check, or what they do with the fact that you get a different answer every time you ask the same question
I came out of every call more confused than when I went in.
For people who are actually doing this, what does the real methodology look like? Specifically: how many prompts are enough to trust the numbers, how often should you be checking, and is anyone actually tying this to pipeline, or are we all just dressing up guesses as data?
Not trying to bash GEO as a whole. Just trying to figure out where the line is between real measurement and the emperor's new dashboard.