The recorded track is dense enough to read each junction, so VM can draw the tulips automatically from the shape of the road. You can fine‑tune any of them afterwards.
Transcribing your voice notes or written pages into directions is a separate AI step — it isn't done here.
Each waypoint's tulip is compared with the turn read from the GPS track. Tap a flagged one to jump to it. Word‑level checks (voice vs written notes) arrive with transcription.