You will read the four reasons the principal recommended pause and the two months of additional clinical-evidence work the trust commissioned instead.
The trust came back to deployment six months later with a smaller scope and a stronger case. The smaller case has now run for two years without a clinical incident.
“If the case is poor, the most expensive thing you can do is to win the engagement.”
We will not pretend AI is the right answer for a problem it does not fit. We have walked away from work where the case was poor. We will publish those notes too.