We ran the trial across our partner clinic in Sydney for 12 weeks — same fourteen GPs, same patient mix, ambient AI scribe enabled half the days, manual typing the other half.
Numbers that surprised us
Across 6,420 consults the ambient-scribe days averaged 3.2 minutes less per consult than the typing days. We expected gains in long-form notes; we did not expect the gains to extend to simple acute visits.
Documentation completeness — measured against the practice's own template — went up 14 percentage points. Doctors had time to write things they were previously skipping.
What patients noticed
“It felt like the doctor was actually listening for once. They weren't typing the whole time.”
Patient, 67, post-consult survey
Patient-rated 'I felt heard' scores moved from 7.4 to 8.6 on a 10-point scale. The mechanism is obvious in hindsight: when the clinician's eyes are on you, not the keyboard, the room reads differently.
What didn't change
Diagnostic accuracy didn't move (it was already high), and total consults per day didn't significantly increase — most clinicians used the recovered time for documentation depth or short between-consult breaks rather than packing in more visits.