A doctor's last appointment might finish at 5:00 PM.
That doesn't necessarily mean their work finishes at 5:00 PM.
There are still consultation notes to complete. Test results to record. Medication changes to document. Referral letters to prepare. Follow-ups to arrange.
And sometimes, the laptop comes home too.
This is one of the less visible problems in healthcare: the consultation ends, but the documentation doesn't.
The Patient Leaves. The Work Doesn't.
Think about a normal consultation.
A patient walks in and explains what's wrong.
The doctor listens, asks questions, checks previous history, performs an examination, discusses possible next steps and creates a plan.
But the important parts of that conversation also need to become part of the medical record.
That could mean documenting:
the patient's presenting complaint relevant medical history symptoms and observations examination findings medications investigations assessment treatment plan referrals follow-up instructions
Multiply that by 15, 20 or more patients in a day and documentation becomes a serious amount of work.
And this is where the problem starts.
Doctors Have Two Jobs Happening at the Same Time
During a consultation, a doctor is trying to do something very human:
listen to the patient.
At the same time, they may also be trying to document what is happening.
So you sometimes get a familiar situation.
The patient is explaining something important.
The doctor is listening.
But they're also looking at a screen and typing.
That's not because the doctor doesn't care about the conversation.
It's because the information has to be captured somewhere.
If it isn't documented during the consultation, it has to be remembered and written afterwards.
Neither option is particularly ideal.
Why Does Documentation Take So Long?
It's easy to think:
"It's only a few notes. How long can that really take?"
But clinical documentation isn't simply writing a summary of a conversation.
The information has to be useful later.
Another clinician may need to understand what happened during the appointment. The record may need to support continuity of care, prescriptions, referrals, investigations, billing or other clinical and administrative processes.
So a doctor isn't simply writing:
Patient has been feeling unwell.
The documentation needs structure and context.
What did the patient report?
What was observed?
What was discussed?
What was decided?
What happens next?
And when that process repeats throughout the day, a few minutes here and there quickly become a significant amount of time.
This Is Where AI Scribes Become Interesting
Now imagine the consultation happening normally.
Doctor and patient talk.
The doctor isn't constantly stopping to type every detail.
With appropriate patient consent and safeguards, an AI medical scribe can capture the conversation and turn the relevant information into a draft clinical note.
The workflow becomes something like this:
Consultation → Conversation captured → Draft note generated → Doctor reviews → Corrections made → Doctor approves
That distinction matters.
The AI isn't there to replace the doctor.
And it shouldn't be deciding what diagnosis the patient has.
Its job is much narrower:
help with the documentation.
Instead of the doctor starting with a blank page after every consultation, there is already a structured draft waiting to be reviewed.
The Blank Page Is Part of the Problem
This is one of the practical advantages of an AI scribe that gets overlooked.
Starting a clinical note from nothing requires effort.
You have to remember the conversation, decide what matters, structure it properly and type it into the system.
Compare that with reviewing a draft.
The doctor can read it and think:
"Correct."
"Remove that."
"That isn't what I meant."
"Add this result."
"Change the assessment."
That's a very different workflow from reconstructing an entire consultation afterwards.
It doesn't eliminate documentation.
It changes where the effort goes.
Less time creating the first draft.
More time reviewing whether the record is actually correct.
But AI Doesn't Magically Remove the Work
This is where a lot of AI marketing gets carried away.
An AI scribe isn't a button that suddenly gives every doctor hours of free time.
The generated note still needs review.
The AI can misunderstand something.
It can miss an important detail.
It can attribute something to the wrong speaker.
It can produce unnecessary information.
And, like other generative AI systems, it can occasionally generate information that wasn't actually said.
So the goal shouldn't be:
"Let AI write the medical record."
It should be:
"Let AI prepare the draft so the doctor can focus on reviewing it."
The doctor stays in control of the final clinical documentation.
Does It Actually Save Time?
There is encouraging evidence, but the numbers need to be kept realistic.
The amount of time saved depends on the doctor, specialty, consultation type, existing workflow and the AI system being used.
Some doctors may see substantial improvements.
Others may save only a few minutes.
And poorly generated notes can actually create additional editing work.
That's why clinics shouldn't buy an AI scribe because a website promises something like:
"Save two hours every day."
The better question is:
How much time does this save in our clinic?
Measure it.
Compare documentation time before and after implementation.
Look at how much editing doctors are doing.
Check whether notes are being completed sooner.
See whether doctors are spending less time catching up after clinic.
Then decide whether the system is actually helping.
The Bigger Opportunity Isn't Just Faster Notes
Saving documentation time is useful.
But there is another benefit that may be even more important.
Attention.
If a doctor doesn't have to spend as much of the consultation thinking about typing everything down, they may be able to focus more naturally on the patient in front of them.
Listen.
Ask the next question.
Notice something important.
Have an actual conversation.
Technology in healthcare shouldn't create another screen demanding attention.
When it's designed properly, it should do the opposite.
It should quietly handle some of the background work.
And Documentation Is Only One Part of the Workflow
This is where we think medical AI becomes much more interesting.
Imagine the consultation ending.
The clinical note is drafted.
The doctor reviews it.
But then the system can also help trigger the administrative work around that consultation.
For example:
Patient Consultation
↓
AI Drafts Clinical Note
↓
Doctor Reviews & Approves
↓
Record Updated
↓
Follow-Up Task Created
↓
Appointment Reminder Sent
↓
Relevant Staff Notified
Now AI isn't just producing text.
It's becoming part of a properly designed clinic workflow.
And that's the kind of medical AI we're interested in building at ELRQ.
Not AI that tries to become the doctor.
AI that handles the repetitive work happening around the doctor.
Because the best use of AI in a clinic may not be making more clinical decisions.
It may simply be giving doctors more of their time back.