What they did
Frank runs a company called Gildre whose YouTube channel had more than 135 videos. Over time he had only managed to comment on 10 or 15 of them himself. He handed the rest to the AI helpers he uses day to day, MUSE and a Grok bot. His instructions covered what kind of thing to say rather than a generic compliment, how cheeky the tone could be, a cap of one or two emojis per comment, and characters and phrasings to avoid so the comments would not read as written by a bot. The job was to comment on each video and give it a thumbs up.
How it went
A week after he set it going, every video on the channel had a comment and a like, which he says read in his own voice. He calls it a big time saver but gives no count of hours, cost or edits. He also reports the helpers going through his inbox to find and cancel his email subscriptions, and finding his name, address and mobile number on 50 websites and emailing each one to ask for removal.
Worth knowing
The post does not say which of the two agents did the YouTube job or how it was signed in to the channel. Frank stresses that he reviews the output before it matters, and that specific instructions on tone and limits were what made the comments sound like him. Mass commenting on your own channel's videos may look odd to viewers if the comments are not clearly from the owner.