Trainer in Austin pointed out my AI chatbot's personality flaw after 10 minutes
I was demoing an AI support bot I built and a UX trainer walked over, said "your bot apologizes before it even knows what went wrong." She was right, it said sorry 8 times in one chat. I rewrote the prompts to drop the groveling and customer ratings went up noticeably, has anyone else found their AI sounds too apologetic?
Is saying sorry 8 times really a problem if customers were still talking to it instead of bailing? Ratings going up after you cut the groveling doesn't prove sorry was the issue, could be a dozen other things changed too.