And you lose quality of service with each move like this. Generic call centers serving a myriad of companies are completely useless to the customer, just read off a script and can't fix anything. AI is essentially the same.
Would anyone whose been keeping up with LLM progress be surprised if within the next 3 years all AI powered call centers are significantly more useful than all of the current call centers?
I worked at Shopify when it was much smaller, it was fairly common for support staff to reach out with questions for the product team which drive fixing bugs or improving features.
How can the AI chatbot fill that role? It’s a gap that offshoring/disintegrating support creates.
It will answer questions that customers have and then it will feedback to the product team about the most common queries made by customers.
Honestly, the Hacker News crowd have been such visionaries about the potential for technology but as soon as that technology infringes upon their livelihood they develop these massive blind spots.
Customer support execs in most companies at least either have documentation or enough experience to understand where the customer is heading with the discussion and resolve problems quicker. Most customers especially larger ones would have account executives who are aware of the customer's needs and personalities and how to deal with them/help them.
I do not think an AI chatbot can replace that, at least as of now. And for those, quoting `it will get better in x years`, do note that those entities who have resources to fund such researches and improve things are slowly evolving from research first to profit first, which means we might not get the AI future we want, but what they want.
This gives me the sense there's a nice training set, and a fairly clear way to measure the AI's performance. "It will get better in X years" sounds very likely
TIL the HN crows works in customer support/call centers!