Not what you think they ask. What they type, in their own words, at midnight, with a diagnosis they do not yet understand. And then: who gets named today when they ask, and why it is not you.
We call this step Context. You can call it listening.
Agencies write first and research afterwards, or research and write at the same time. That is how a practice ends up with forty articles answering questions nobody is asking, and a monthly report full of impressions instead of patients.
This is also where we find out whether we should take the work at all. If the demand in your specialty is decided by how close your building is to the patient, then no amount of writing moves it, and we will say so before you spend anything.
You get content. It is just content about the wrong things, and you will not find out for six months.
query set extraction and intent classification · AI Overview and chatbot citation mapping across four engines · competitor entity audits · local-pack versus content-decided SERP analysis · baseline citation report
The free audit tells you exactly that — which of the six are done, which are half done, and which have never been started.