The visit note, drafted while you see the patient
Sona hears Malay, English and Chinese. Nothing is filed until the doctor signs.
Selamat pagi, Puan Nur Aisyah. Apa yang boleh saya bantu hari ini?
Doktor, saya sakit bila telan. Dah 3 hari susah nak makan.It hurts to swallow. Hard to eat for 3 days.
Ada demam, batuk atau selsema?
Semalam demam sikit, hari ini dah okey. Saya makan panadol 2 kali semalam.
Scribe drafted this notefor you to sign
Presenting complaint
- Doktor, saya sakit bila telan.0:07
- Dah 3 hari susah nak makan.0:07
- Semalam demam sikit, hari ini dah okey.0:15
- Saya makan panadol 2 kali semalam.0:15
- Tak ada alahan, doktor.0:27
Signing files the note. The recording is deleted from this iPhone.
Rebuilt from Sona Pro on iPhone.
In the room
Nobody types while the doctor examines
The patient agrees, and Sona Pro records the visit on the doctor's iPhone. It tells the voices apart as they speak, keeps Malay and Chinese as they were said, and the doctor names each voice with one tap.
- Malay, English and Chinese in one consultation, and Tamil mixed in.
- Care gaps show at the top, such as a screening not on record.
- The note itself is written in English.
Sources
Tap a line to find it in the visit
Every statement in the draft carries the moment it was said. Choose one and the transcript jumps to it, so the doctor checks the source before signing, not after.
Presenting complaint
- Voice 1, 0:02
Selamat pagi, Puan Nur Aisyah. Apa yang boleh saya bantu hari ini?
What the model saw: Selamat pagi, Puan Patient name hidden. Apa yang boleh saya bantu hari ini?
- Voice 2, 0:07
Doktor, saya sakit bila telan. Dah 3 hari susah nak makan.
- Voice 1, 0:12
Ada demam, batuk atau selsema?
- Voice 2, 0:15
Semalam demam sikit, hari ini dah okey. Tak batuk, cuma hidung tersumbat. Saya makan panadol 2 kali semalam.
- Voice 1, 0:24
Ada alahan pada apa-apa ubat?
- Voice 2, 0:27
Tak ada alahan, doktor.
Unsourced lines
A line with no source waits for the doctor
Each statement links to what it came from. One with no source stays out of the note unless the doctor adds it, and the note cannot be signed until that is decided. Decide it here.
- 1Check in10:45
- 2Rooming and vitalsDone 11:03
- 3Visit note1 line to decide
- 4OrdersNone yet
- 5Sign and closeOne passkey
Before you sign
Nur Aisyah binti Rahman, today 11:00- Note written
- Diagnosis set
- Allergies checked
- Every line sourced or decided
Sources
Statement coverage 4 of 5 sourced
- Pain on swallowing for 3 days.0:07
- Mild fever yesterday, none today.0:15
- Took paracetamol twice yesterday.0:15
Symptoms began after a cold last week.No source in the visit. It stays out unless you add it.
Added by Dr Arvind- No known drug allergies.0:27
Rebuilt from Sona Pro on the web.
Privacy
Names and numbers come out first
Names, IC numbers and phone numbers are removed before any text reaches the AI model, which sees only a marker such as Patient name hidden. The recording is turned into text and then deleted; the note and the transcript stay in the record.
3 of 4 details hidden
- PatientNur Aisyah binti Rahman
- IC920314-19-0012
- Born14/03/1992
- Phone+60200000001
- ComplaintSore throat, runny nose and cough for 3 days.
- AssessmentAcute upper respiratory tract infection (J06.9)
Rebuilt from a real run of the privacy filter in Sona, 12 Nov 2026. Drag the line.
More from Sona
Patient-facing
- Voice AI agents
Sona picks up in Malay, English and Chinese and prepares the booking for your front desk.
- Intake & triage
Emergency words in four languages get the 999 advice and a care-team task.
- Sona Care
Visits, results and questions in plain words, behind an identity check.
- Replies in their language
Written replies in Tamil, Malay, Chinese or English.
Provider-facing
- Clinical scribe & copilot
Drafts the note during the visit, for the doctor to sign.
You are here - Sona Pro
What needs your staff first, on the web and on iPhone.
- One patient record
Calls, chats, visits and claims write to the same patient record.
- Staff approvals
Money and the clinical record wait for a person to approve.
- Health world model
What is most likely next for a patient, counted out of 100 futures, for a clinician to weigh.



