It implies that operational losses tend to follow a normal distribution.

It is more effective in modeling some types of operational risk, such as losses from fraud, than others, such as losses from natural disasters.

It is generally used to estimate routine operational losses which occur at a relatively high frequency.

It is suitable for modeling the tail of the operational loss distribution, but not for modeling the body of the distribution.