AI Visit Summary in Telehealth

Live transcription during a call

During a telehealth call, on the practitioner's side, clicking Transcribe session starts live speech-to-text, shown as a "Live transcript" strip as the call happens.

Live transcription isn't itself an AI feature The live transcript is produced by the browser's own built-in speech recognition, running locally in the call window. It's not a cloud or AI service on its own — the AI step comes later, when the call ends and the transcript is summarized.

What you'll learn

How to start live transcription during a telehealth call
What the AI-generated Session Summary contains and how it's labeled
Why editing the summary before saving is a required step, not optional polish
What happens if no transcript exists or the summary can't be generated

An active telehealth call with the Live transcript strip visible at the bottom of the practitioner's screen after clicking Transcribe session.

When the doctor ends the call, a Session Summary modal opens automatically. It's explicitly labeled "AI-generated · [duration]" — this is the clearest, most explicit AI disclosure anywhere in ClinyPal.

The modal shows a structured clinical summary built from the call transcript, organized into four sections:

Chief Complaint
Findings
Plan
Follow-up

Below the structured summary, a collapsible section holds the raw transcript, so you can check the AI's summary against exactly what was said.

AI-generated, explicitly labeled The Session Summary modal is labeled "AI-generated · [duration]" directly in its header — the most explicit AI disclosure in ClinyPal. It's built entirely from the call's transcript, not from any other part of the patient's record.

Edit-then-save: a required review gate

Before the summary can be saved, the doctor edits it — correcting anything the AI got wrong, adding detail it missed, or rephrasing clinical language. This mandatory edit-then-save step acts as a human-review gate: the AI's draft doesn't reach the patient's record as-is, it reaches it after a clinician has reviewed and confirmed it.

End the call

The Session Summary modal opens automatically.

Review the structured summary

Read through Chief Complaint, Findings, Plan, and Follow-up.

Check against the transcript

Expand the collapsible raw transcript if you want to verify a detail.

Edit the summary

Correct or expand the text as needed — this step happens before saving.

Click Save to EHR

The reviewed, edited summary is saved to the patient's record.

Review before saving to EHR The AI's draft summary is a starting point for clinical documentation, not a final note. Review it against the transcript and your own memory of the call before saving.

When something goes wrong

The fail-soft messaging here is unusually explicit, because clinical documentation is at stake. If no transcript was recorded, the AI call fails, or the summary service is unreachable, the modal says so clearly and tells the doctor to write their own notes from the raw transcript instead — nothing is silently hidden or left blank without explanation.

No transcript recorded — the modal states this plainly rather than showing an empty summary.
The AI call fails — the modal tells you the summary couldn't be generated and to write your own notes.
The summary service is unreachable — same clear messaging, with the raw transcript (if one exists) still available to work from.

Running a Telehealth Session

A full walkthrough of starting, running, and ending a telehealth call, including where Transcribe session and the Session Summary fit into the flow.