The promise was simple.
Brief the AI once.
Let it work.
Give it a light review.
Ten percent in.
Eighty the model.
Ten the check.
That was the pitch.
Here is the ratio most clinicians live.
Fifty percent setting up the input.
Ten percent the output that touches the patient.
Forty percent checking it.
Correcting it.
Carrying it.
The load did not disappear.
It moved rooms.
Every recommendation still needs the clinician.
At the front.
And at the back.
They called it delegation.
It worked like routing friction
with a recommendation attached.
Count the week honestly
and the architecture is visible.
Not because anything broke.
Because the measurement showed
what the institution had been buying around.
The problem was never capacity.
It was the architecture upstream of the call.
↳ The Governance Owner
names what the agent is authorized to do.
↳ The Decision Owner
holds the clinical call.
↳ The Handoff between them
is where the gap closes.
The Accountability Gap™ (TAG™)
does not close when you add better models.
It closes when you fix the decision
before it reaches the bedside.
You already know your real ratio.
The only question is whether the seat was named
before the next deployment goes live.
Mo Johnson, MD MBA is a cardiothoracic surgeon and the founder of GPe Research. Field Notes are short dispatches from the clinical AI accountability frontier, published alongside the MedicoVigilance™ newsletter at medicovigilance.org.
Follow the work on LinkedIn: linkedin.com/in/mo-johnson