CASE STUDY · HEALTHCARE AI & CLINICAL WORKFLOW AUTOMATION
Cutting physician charting time by 35% with AI-assisted documentation.
- 35%Charting time down from 120 to 78 minutes a day (35% less) in a 6-week pilot with 12 clinicians
01 · PROJECT OVERVIEW
The client and the brief.
An anonymised mid-market clinic group. Name withheld under confidentiality.
A clinician-controlled drafting workflow integrated with the EHR: the AI drafts, the clinician edits and signs, and every output is logged.
- 1Clinicians spent about 2 hours a day on charting, often after hours.
- 2Notes were written from memory at the end of the day, or at home in the evening.
- 3Clinicians described this as the part of the job they liked least.
- 4Management saw the same problem from the other side: notes completed late meant claims sent late, and handovers between clinicians depended on notes that were sometimes thin.
- 5The group had looked at AI documentation tools before.
02 · WHAT WENT WRONG
What went wrong, and how we fixed it.
Every engagement hits problems. These were the ones that mattered here, and what we changed so they did not come back.
- PROBLEM 01Early drafts mixed up similar medications and sometimes put details under the wrong heading
Clinicians caught them, which is the reason the sign-off step exists, but trust dipped in week two. We changed three things. Medication names are now taken from the structured medication list in the EHR and not from free text.
- PROBLEM 02Sections are filled from tagged parts of the conversation
And any medication or dose in the draft is highlighted for extra attention. The error reports from clinicians dropped and use went up.
- PROBLEM 03A second hurdle was the consent and privacy side
Some clinicians were unsure when it was appropriate to capture a consultation. We worked with the medical director and the compliance lead on a short patient notice and a rule that the clinician can switch capture off at any point. Visit types involving sensitive topics were left out of the pilot by default.
- PROBLEM 04This took a week to settle and was worth it, because it removed the clinicians' main worry
03 · THE PROCESS
How we ran it.
A 6-week pilot with 12 clinicians.
- 01Week 1
we shadowed clinicians, measured charting time with a short daily log and agreed a note format for each visit type with the medical director.
- 02Weeks 2 to 3
we built the drafting workflow and tested it with two clinicians on real but non-critical visit types. They reviewed every draft and marked errors by type.
- 03Weeks 4 to 6
we widened to all 12 clinicians. We held a short weekly review with them, fixed the most common errors and watched time saved.
The medical director attended every review. Clinicians could ask for a visit type to be removed from the pilot at any time.
04 · RESULTS
What changed.
Measured against the way the work ran before the engagement began.
- NOW 01
Clinicians review and sign a draft instead of writing from scratch, and more notes are finished during the working day.
- NOW 02
The medical director reviews the edit report each month to decide where the drafts need improving.
- NOW 03
New visit types are added only after a short trial with the clinicians who see them.
- 35%
Charting time down from 120 to 78 minutes a day (35% less) in a 6-week pilot with 12 clinicians
05 · WHAT WE BUILT
How it works.
- The consultation content is captured with the clinician's knowledge, turned into text and sent to a drafting service that produces a note in the group's agreed format.
- The draft appears inside the EHR as unsigned.Nothing is saved to the record until the clinician signs.
- The integration uses the EHR's standard interfaces.Data sent to the drafting service is limited to what the note needs.
- Access is by role.Each draft is logged with the input, the output, the edits the clinician made and the sign-off time.
- A review report shows which sections clinicians edit most, which tells us where the drafts are weakest.
- Handover
- We trained the clinic's super-users on the edit report and left a one-page guide on what to do when the drafting service is down: clinicians simply write the note as before, and nothing blocks the visit.
06 · QUICK ANSWERS
The short version.
- What problem did the client have?
- Clinicians spent about 2 hours a day on charting, often after hours.
- What did Brandsmashers build?
- A clinician-controlled drafting workflow integrated with the EHR: the AI drafts, the clinician edits and signs, and every output is logged.
- How was the work run?
- A 6-week pilot with 12 clinicians.
- What were the results?
- Charting time down from 120 to 78 minutes a day (35% less) in a 6-week pilot with 12 clinicians.
YOUR TURN
Have a similar requirement?
Tell us what is stuck. We will come back with a plan and matched, pre-vetted engineers.
