When you ask an AI a question, it often doesn't run one search. It fans the query out into several sub-queries, runs them in parallel, and synthesises the combined results into one answer. Ask "best fragrance-free moisturizer for eczema under $30," and behind the scenes the engine might search for "fragrance-free moisturizer eczema," "best moisturizer sensitive skin," "affordable eczema moisturizer," and more.
This changes the visibility game in two ways. First, you don't have to match the user's exact phrasing. You have to be retrievable for the underlying sub-questions, which broadens the surface you're competing on. Second, a single answer can pull from many sources at once, so breadth of relevant, well-structured content matters more than a single perfectly-optimised page.
It's why "the questions that matter" for a brand number in the hundreds, and why measuring AI visibility means sampling across many prompts, not checking one.