Skip to content
Curious Context Labs

case study · Healthcare · Custom app

Giving clinicians their evenings back from note-writing

How an AI scribe can turn a consented consultation into a draft note that the clinician checks and signs, without changing who owns the record.

business
Healthcare clinics
build
Custom app
timeline
8–10 weeks
read
3 min

01 the business

Multi-specialty outpatient group

140 clinicians across 9 clinics

Clinicians see 20 or more patients a day and finish their notes at home. They type during visits and patients notice. Two clinicians have cut their hours, citing documentation as the main reason.

02 before

The problem

Clinicians type while the patient talks, then finish their notes at night. The record gets the clinician's attention that the patient should have had, and documentation is one of the biggest drivers of burnout. Human scribes help but are expensive and hard to find.

  • Notes are finished after hours, often from memory.
  • Clinicians face the screen instead of the patient.
  • Note quality and length vary widely between clinicians.

03 the build

What the AI does

  1. Step 1: Asks for and records the patient's consent at the start of the visit, and stops recording if either person says no.
  2. Step 2: Listens to the consultation on a phone or computer and writes a transcript as it goes.
  3. Step 3: Drafts the note in your specialty's template, such as history, examination, assessment and plan, using only what was said.
  4. Step 4: Marks anything it was unsure about, such as a medication name or dose, for the clinician to check.
  5. Step 5: Places the draft into Epic, Oracle Health or your practice system as an unsigned note, and deletes the audio after the agreed period.

paused: waiting for a person to approve

How it was put together

  • Chose a scribe that records only with consent and deletes audio after an agreed period.
  • Built note templates with each specialty's lead clinician.
  • Placed drafts in the record as unsigned notes, with unclear terms such as drug names flagged.
  • Signed business associate and data processing agreements before any patient data flowed.

04 human in the loop

Where a person stays in charge

The clinician reads, edits and signs every note; nothing is filed without their signature. The tool writes down what was said and never suggests a diagnosis or treatment.

  • Patients are asked for consent at every visit and can say no without any effect on their care.
  • Clinicians read, edit and sign every note; nothing is filed unsigned.
  • The scribe documents what was said and never suggests a diagnosis or treatment.

05 rollout

From kickoff to live in 8–10 weeks

  1. Weeks 1–3Privacy and security review, consent wording, contracts and template design.
  2. Weeks 4–6Pilot with 15 volunteer clinicians; review a sample of notes for accuracy each week.
  3. Weeks 7–10Extend clinic by clinic, with a named clinical lead and ongoing note audits.

06 after

What changes

Clinicians look at the patient instead of the screen, and notes are finished before the next patient instead of at night. The note still belongs to the clinician, who checks every line.

worked example · your numbers will differ

Assume 140 clinicians each spending 60 minutes a day finishing notes after clinic, and the scribe cutting that by a third, over 220 clinic days.

140 × 20 minutes × 220 days = 616,000 minutes, or about 10,270 hours a year returned to clinicians.

What to measure

  • Time spent in notes per visit and after hours.
  • Share of notes signed the same day.
  • Errors found in note audits, by type.
  • Clinician and patient experience scores.

07 evidence

What the research says

  1. Observed physicians spent 27% of their office day on direct time with patients and 49.2% on the electronic record and desk work, plus 1 to 2 hours of after-hours work each night.

    Annals of Internal Medicine, Allocation of physician time in ambulatory practice (Sinsky et al.) (opens in a new tab) · 2016

  2. In the first 10 weeks of a rollout, 3,442 physicians used ambient AI scribes in up to 303,266 patient encounters; use was linked with less time on documentation, though accuracy still needs a clinician's review.

    NEJM Catalyst / Kaiser Permanente, Ambient artificial intelligence scribes to alleviate the burden of clinical documentation (Tierney et al.) (opens in a new tab) · 2024

  3. 43.2% of US physicians reported at least one symptom of burnout in 2024, down from 48.2% in 2023 and 53% in 2022.

    American Medical Association, U.S. physician burnout hits lowest rate since COVID-19 (opens in a new tab) · 2025

08 lessons

Mistakes to avoid

  • Skipping consent. Patients must be asked, every time.
  • Signing without reading. Scribes can mishear names and doses.
  • Ignoring specialties. A generic template frustrates everyone.

09 questions

Common questions

Do AI scribes work at scale?
Kaiser Permanente reported 3,442 physicians using ambient scribes in up to 303,266 encounters within 10 weeks, linked with less documentation time, with accuracy still needing review.
Is the audio kept?
Only for the period you agree, usually long enough to finish and sign the note, then it is deleted.

keep reading

All case studies ->

Tell us how your business runs. We'll tell you where AI fits.

A free 30-minute call. You talk, we ask questions, and you leave with a few concrete ideas — or an honest answer that AI isn't the right fix yet.

Working with teams around the world.