While building Cited I've been asking ChatGPT and Claude the questions real buyers ask before choosing software : "best CRM for a small sales team," "Mixpanel alternatives for startups," that kind of thing.

Two patterns surprised me.
First, the big names only dominate the generic prompts i.e the moment a buyer adds a use case, region, or company size ("expense tool for a 50-person team in the UAE," "CRM for founder-led sales"), the answers change completely, and different names surface. Most teams have no idea which of these buying prompts they win, which they lose, or why.

Second, the answers aren't static : they shift as the models pick up new content from comparisons, reviews, and docs. Which means the gaps are actually closeable : track the prompts that matter for you, publish the content that fills the holes, and your visibility genuinely moves week over week.

1. Curious about two things from folks here:

   1. Have you checked what AI says about your product or category, especially on the specific prompts your buyers would ask? Anything surprise you?

   2. If you're actively trying to improve how AI answers about you : what's actually worked?

   We launch Thursday (16th July), happy to share what we've learned from auditing a few hundred B2B categories in return.