Prompt Discovery

Prompt discovery combines several sources. First, expand your existing query bank with semantic variations, long-tail extensions, and intent-modified versions of the prompts you track. Second, read AI answers and work backwards to the prompts that would produce responses mentioning competitors but not you. Third, use real customer language: sales call notes, support tickets, community threads, and survey answers.

Deduplicate the prompts and cluster them into prompt families, meaning groups of semantically similar prompts. Then rank each family on an opportunity score that considers likely frequency, commercial intent, your current absence, competitor presence, and whether you can create content that closes the gap.

Review the ranked list by opportunity, prompt family, intent type (informational, commercial, navigational), and competitive density. For each prompt, check which engines answer it, which competitors appear, and what content would answer it best.