One of the pitfalls of automated warnings (regardless of what they are based on) is they tend to “hold on” to warnings too long, since they cannot be expired early. Here is an example where a cell had been weakening for some time (>20 min), yet the 45 min Dangerous Thunderstorm Alert (DTA) from ENTLN was still valid. I imagine a forecaster in this situation would expire the warning long before the DTA expired.

Ertel