Volume was always a proxy
Keyword tools estimate how many people type a phrase. What you actually wanted to know is which questions precede a purchase in your category. Those overlap, but the highest-volume terms are usually the least commercial, and chasing them produces traffic that never buys.
Start with the questions you already receive
Sales calls, support tickets, and the messages people send before signing are a record of real pre-purchase questions in buyers’ own words. That is the same source that makes an FAQ page work again, and it costs nothing to mine.
Sort by where the question sits in the decision
Questions cluster into three groups: what is this category, which option should I pick, and how do I implement it. Middle-of-decision questions convert best and are the least covered, because they require an opinion competitors avoid publishing.
The best question to answer is the one your competitors are too cautious to answer honestly.
Test the prompt, not the keyword
Ask the assistants the question as a buyer would phrase it and read the answer. You learn immediately whether an answer already exists, who is being cited, and what is missing — which is more informative than any rank position and takes minutes.
Prefer questions you can answer specifically
A question where you can give a number, a threshold, or a rule you actually follow is worth ten where you can only generalise. Specificity is what makes a passage quotable, and generic coverage of a popular question adds nothing to what a model already knows.
Then stop expanding and go deeper
Once the genuine questions are covered, more pages produce diminishing returns and start to look like padding. Improving the answers you have beats manufacturing questions nobody asked — the reasoning behind not building a page per question.