When we started designing Notovera's documentation engine, the first rule we wrote down wasn't about speed. It was that no note is saved until a therapist has reviewed and signed it. Not when the draft looks perfect, not when nothing is flagged. Every time.
That might sound like it slows things down. In practice it's the opposite. When you know the draft can't go anywhere without you, you read it the way you'd read a colleague's note: quickly, looking for what's off. The AI does the typing; you do the judging.
The same principle shapes goals. There's no fixed goal bank in Notovera. Goals are drafted from what you found at evaluation, and you approve, edit or replace them. At re-evaluation the draft summarizes progress from prior notes, and you bring the updated assessment.
We think this is the only responsible way to put AI into clinical documentation. It's also the only way we'd want to use it ourselves.