Why LLMs like ChatGPT are bad for customer service
crisp.chat
2 threads
The insights on "few-shot" learnings are very interesting. We also wanted to use a vector db combined with GPT for our search engine. Intuitively, I would have thought the accuracy with this method to be much higher. Models hallucinating or giving wrong answers +20% of the time is not possible for direct answers, as helpers it's a good idea!
Wondering if there would be a way to create a score of confidence to the answers from this method? And discard the lowest ones.
Love the part about why ChatGPT is bad for customer service. Some many app builders are doing it wrong nowadays.
The best support on earth is when customers don't even need support, and it's the main reason why LLMs won't solve this problem :)