Proprietary GPT instance · ChatGPT Users

Proprietary GPT instance

My client is a very large NY based catering company. They want an LLM that will give their sales people the ability to ideate new proposals and offers to clients, drawing from an existing catering database (their own), Salesforce data, a recipe database and maybe from some other sources. I think they need a vector database (Pinecone?) that pulls from all the sources mentioned and through embeddings would learn to answer questions from this information. Am I doing this the right way? Anybody have any experience with similar situations?

1 comment