My boss said AI would replace our QA team, 8 months later he was half right

Back in March my manager at a call center in Phoenix told us all to stop fighting the new AI monitoring tool because "it's going to do your jobs better than you by summer." I kept logging my own notes on every call anyway, just in case. The AI flagged 40% of our calls as needing follow-up, but when we actually listened to the flagged ones, over half were false alarms from background noise or overlapping voices. Meanwhile it missed a real compliance issue that I caught by ear because the customer had a thick accent the AI kept misreading. Now they're keeping us but we have to double-check every single AI flag, which takes longer than just doing it ourselves. Has anyone else had to babysit an AI tool that was supposed to replace them?

2 comments

2 Comments

simon71721d ago

Does your AI tool also flag every time someone sighs or coughs near the mic? Ours thought a customer was saying "cancel" when she was actually asking about our "canvas" print options. We spent a whole afternoon calling people back who never complained about anything. The funniest part is the AI logs are now considered the "official record" but we all know they're basically guessing. I swear, half the time I feel like I'm teaching a toddler to use a phone.

hugo_schmidt20d agoMost Upvoted

thats hilarious but honestly not surprising. our system once flagged a lady for saying "bullshit" when she was actually asking about our "bulk" shipping rates. we sent her a formal apology email and everything. turns out she had a heavy accent and the AI just heard what it wanted to. now our team makes bingo cards out of the worst transcriptions during lunch breaks. the best one so far was a guy who supposedly demanded "a million free tacos" but he was actually ordering a single bean burrito.