Question

A financial services company is building an internal chatbot using a generative AI model to answer employee questions about company policies. During testing, the model sometimes fabricates policy details that do not exist in the official documents. The team wants the model's answers to be based on the company's actual policy documents to improve accuracy. Which approach best addresses this problem?

Accepted Answer

Ground the model's responses in the company's policy documents using a retrieval-based approach such as Vertex AI Search

Answer

Increase the model's temperature setting so it produces more creative and varied answers

Answer

Switch to a pre-trained vision API to analyze the policy documents as images

Answer

Deploy the chatbot on more powerful GPUs to reduce the frequency of incorrect answers