I'm not 100% sure this is what you're asking but it should only be set to auto-delete exactly that phrase when it exists on its own because that phrase can exist in a proper response. It should also not delete any reply that begins with the phrase but has anything extra on the end. Granted these are easily abused, but it comes down to a choice between preventing abuse and possibly losing legitimate replies or allowing a little legroom and not losing anything potentially helpful.
Rules should only be applied to the absolute minimum possible.
Suppose someone asked a question and the reply responds with the answer but is also offering to help the person see what they were doing wrong or help them on a different thing they were trying, it could easily look something like:
"I want to block routing within my switches but only for specific
vlans, but it doesn't seem to be working, what's something simple I can use to block traffic between vlans"
Then the reply comes in
"ACLs, what have you tried?"
Here a rule stating that adding "a little more" to the phrase should also be deleted would probably lead the rule to delete this reply which did in fact give a proper response but foolishly used a phrase that has caused issues/concerns in the past.
Keeping it to only delete a reply that matches that specific phrase means that you know exactly what you're deleting and will never accidentally delete a reply that was actually trying to be helpful.
The rule should be "delete exact matches, flag possible transgressions"
Then later (maybe your 5 hours) someone can check it, and say, "yeah, let's delete that" or "ah, that looks ok, let's not delete that one"