One is the type where people are working to actually help the customer and the rep is authorized to take whatever action necessary to accomplish that goal.
The other is the type where the rep has no authority to do anything. They'll endlessly transfer the caller to another rep, make them wait, transfer again, make up excuses for why nothing can be done, and just do what they can to goad the caller into consenting to giving up, absolving the company of having to pay anything or provide any sort of service.
I don't think number one will be replaced by AI. I think number two will be replaced by AI.
But it doesn't really matter what "people" think. Corporate sees a way to save money by employing less people, customers be damned. Sadly customer service does not seem to have any bearing on a company's success, considering Google is successful as it is when it literally does not offer any kind of customer service at all.
Lex was the state of the art for these sorts of systems at the time.
These newer LLM based systems won't be perfect, but they'll be 10x better at free form intent recognition, and significantly less repetitive and robotic for menus and similar. I suspect the agent experience when you get to a human will be far more comprehensive, enabling real human agents to be just a little bit more helpful sounding.
Realistically though, anything that could be used to improve agent or callers experience in these environments will be squeezed and stretched in the name of profit to the point where both parties suffer.
Yes.
One of the things that I learned, early in my career, is that the best customer service, is good product design, obviating the need for customer service. That’s something within my control.
For many companies, post-sale customer service is a money sink. I remember being told that the margins on a product were so thin, that a single call from a user would wipe out the profit.
But I was also never given support to produce a decent UX, because that was expensive.
Damned if you do[n’t].