71 percent would reject it.
A model that is 99 percent accurate,
rejected because it cannot explain itself.
2,275 finance leaders.
Across North America and EMEA.
IDC. February 2026.
Accuracy stopped being the product.
The explanation is.
Finance teams now spend
nearly 13 hours every week
reconstructing what the model did.
Nearly half spend 15 hours or more.
One in five spend 30.
That is not analysis.
That is people rebuilding
an audit trail nobody built.
The model cannot explain itself.
Often the people who built it
cannot fully explain it either.
So the explanation gets assembled
after the fact,
by whoever will be asked.
What institutions call verification.
↳ Checking the output is correct
↳ Confirming the model performed
↳ Documenting the result
What those 13 hours actually are.
↳ Reconstructing a decision nobody logged
↳ Defending a call nobody authorized
↳ Absorbing exposure nobody assigned
The hours are not the cost.
The hours are the symptom.
Somebody is already doing
the accountability work.
They are doing it unnamed,
after the decision,
without authority to change it.
The Governance Owner names
what the model is allowed to decide.
The Decision Owner holds
what the output becomes.
The Handoff is where the explanation
gets built instead of reconstructed.
The Accountability Gap™ (TAG™)
does not show up as a failure.
It shows up as 13 hours a week.
You already know who in your institution
is doing that reconstruction right now.
That person is not your verification layer.
They are your unnamed owner.
Who would answer for the model
if the regulator asked tomorrow?
Mo Johnson, MD MBA is a cardiothoracic surgeon and the founder of GPe Research. Field Notes are short dispatches from the financial AI accountability frontier, published alongside the FinVigilance™ newsletter at finvigilance.org
Follow the work on LinkedIn: linkedin.com/in/mo-johnson