So the approach is to write a bot, do zero human checks and then say "well, see, yes, we sent this complaint, but we didn't actually send it ourselves, the bot did, so we didn't know that it was wrong, because we've decided to never check these things before sending them out"?

It's weird, when you can write a program to do something in your name and as your agent, and then claim "it totally wasn't me, lol".