Wednesday, September 30, 2026

The end of human customer support?

Those of you who read this blog regularly know how much I can get frustrated with customer service support over the phone. In the US, that kind of service is rarely done domestically, but comes increasingly from the Philippines. 

There, these people mean well and for the most part are grateful to have that job and do the very best they can. 

Depending on their tenure on the job, they know little about what they’re helping callers with and depending on their accent and command of English, they can be hard, yet alone impossible to comprehend so, overall, that’s not a good way to start. 

So, recently, when my call ended up in the Philippines, couldn’t get resolved, needed to be escalated, or depended on someone more qualified to give me a call when they’d be available, I’ve been turning to AI and every time I needed my questions answered or my problem resolved. 

Bottom line, my questions were addressed well beyond my expectations and I got them in writing. So far, it has beat the “human touch”, and with the constant improvements that I’ve observed around AI, I don’t see why this wouldn’t, on the contrary, it’s likely to get even better. 

Now for the bad news; in a not-so-distant future, we can bet that this will be the beginning of the end of these customer support centers staffed with real humans either in the US, in the Philippines, or anywhere else...

2 comments:

오피셜메일 said...

AI로 문의했을 때 답변이 글로 남아 같은 문제를 다시 설명하지 않아도 됐다는 대목이 인상적입니다. 다만 답을 못 찾은 예외나 환불처럼 책임이 따르는 결정은 사람 담당자에게 질문·AI 답변·미해결 사항을 한꺼번에 넘기는 기준도 필요하겠어요. 그러면 응대 속도와 해결 책임을 함께 볼 수 있을 것 같습니다.

Anonymous said...

It is impressive that inquiries handled by AI leave a written record, eliminating the need to re-explain the same issue. However, for exceptional cases where no answer is found or for decisions involving liability—such as refunds—there needs to be a standard procedure for handing over the original inquiry, the AI’s response, and the unresolved details to a human representative all at once. This approach would allow for a balance between rapid response times and accountability for resolving the issue.