The consensus here is the 'ban process' is in 3 stages.
1) Automated detection of bot activity / report by another player
2) Analysis of account by trained moderator using SC developed tools (SC has already stated this)
3) Account is placed on a 'banned list' and will be ban in a wave. The purposed of this is to prevent the bot developers from to trying to deduce the method of detection.
IMHO once a bot account is at stage 2 it is dead, it maybe banned in the next wave or in a wave six months from now. The only condition when a bot account may not be ban is if there are too many accounts for the moderators to process and the account is towards the end of the list.
What is being discussed here is stage 1. ShaneOss's research shows SC has contracted the automated bot detection to a third party and if the third party does not do get the data it needs to analysis the account then the account would never get to stage 2. The counter argument is if the third party does not get the data your account will be automatically placed in stage 2 as only a bot account would try to hide the data.