14
Vent: A customer's feedback on my AI chatbot's tone made me redo everything
I run a small business using an AI chatbot for customer support. Last month, a guy responded to our survey saying the bot sounded like a "cheerful robot who never heard a real person talk." He said it used too many exclamation points and words like "fantastic" and "wonderful." At first I was annoyed, but he had screenshots showing how fake it sounded. So I spent 2 weeks rewriting all 50 of the bot's response templates. I cut the happy talk and made it sound more plain and direct. Now we ask users if it feels more human, and so far 8 out of 10 say yes. Has anyone else had to tone down their AI's personality after getting pushback?
3 comments
Log in to join the discussion
Log In3 Comments
the_kevin20d ago
Those 8 out of 10 people saying yes is a solid proof you were on the right track. It's funny how one blunt customer can save you from two weeks of rewriting fifty templates if you just listen sooner.
5
oscar_ellis20d ago
That thing about "one blunt customer can save you from two weeks of rewriting" really hit home. My buddy runs a little coffee shop and his chatbot kept calling customers "champion" and "legend" in every reply. One guy finally snapped and said "stop trying to mate with me through text." He was so mad he sent a whole paragraph about how it felt like the bot was trying too hard to be his best friend. My friend changed it all to just "sure thing" and "no problem" and customer complaints dropped like crazy. Sometimes you need that one person to tell you the truth straight up.
1