A brief is a client's best account of what they think they need. It is written after the internal argument has already been settled, and it carries the settlement rather than the argument. That makes it useful and incomplete in the same breath.
A survey asks a different question. Not what do you want built, but how does this get described in three different rooms – the room where the work gets done, the room where it gets sold, and the room where it gets funded. The gap between those three descriptions is almost always where the real problem lives.
What the survey actually collects. How the work is described internally. How it is described to customers. How it is described to a board or a council. What has been tried before and quietly abandoned. Where the proof currently lives, and who can reach it.
None of that is exotic. What makes it worth a phase of its own is that nobody has usually written it down in one place, and the act of doing so tends to settle arguments that had been running for years.
What it changes
The map is what everything after checks itself against. It is the reason a decision made in month nine still agrees with one made in month one – and the thing we return to when the work threatens to scatter.