Why does the "AI won't cite for" data sometimes show topics that look unrelated to us?
Updated June 30, 2026 · Reviewed by [email protected]
That list is built from the prompts you are actually tracking, so it can only ever be as tightly scoped as your prompt set. If a prompt's phrasing is broad or adjacent to your category, the model's real answer to it can pull in sources that look tangential even though the prompt itself is one you added.
If a specific result genuinely doesn't belong, the fix is usually the prompt, not the gap detection: tighten or remove the prompt that is producing the off-topic result, or add a more precise version of the question your buyers actually ask. The gap list will follow whatever prompts feed it.