How Does It Work

Our message automation services are based on newest version of ChatGPT and Retrieval-augmented generation (RAG). It utilizes power of modern generative AI, as well as seller's business context, to compose messages as if written by seller's own hands.

Newest version of ChatGPT is astonishingly powerful. It can even correctly solve Math Olympics geometry problems. It definitely has potential to be your message assistant. But apparently we cannot directly throw buyer's messages and let ChatGPT provide an answer, because ChatGPT was trained on general data. It doesn't know about your business, or how you handle customer's questions. That's why RAG has its place. It's a is a technique for enhancing the accuracy and reliability of generative AI models with facts fetched from external sources. In this case, before we asking ChatGPT to answer one buyer's message, we tell ChatGPT details about that message's related item, and for same kind of item how you handle buyer messages previously. This way, ChatGPT is able to compose a reply message as if it's handled by seller's own mind.