Question

A retail company deploys a generative AI chatbot to answer customer questions about products. Their product catalog, pricing, and availability change several times per day. The team wants the chatbot to always reference the most current catalog information without retraining or fine-tuning the model each time data changes. Which approach best meets this requirement?

Accepted Answer

Use retrieval-augmented generation (RAG) to pull current catalog data from a live data source at query time

Answer

Fine-tune the base model nightly on the latest catalog export

Answer

Increase the model's maximum token limit so it can hold the entire catalog in each prompt

Answer

Rely on the pretrained model's built-in knowledge and add a disclaimer about possible outdated information