Question

A retail company wants its gen AI assistant to consistently write product descriptions in the company's distinctive brand voice and tone. They have thousands of examples of past descriptions that reflect this style. The factual accuracy of product specs is handled separately. Which Vertex AI approach best achieves the consistent brand voice?

Accepted Answer

Fine-tune (tune) a foundation model using the company's past product descriptions as training examples

Answer

Ground the model against a data store containing the company's product spec sheets

Answer

Use a pre-built Natural Language API to analyze sentiment in existing descriptions

Answer

Increase the model's temperature setting so it produces more varied, creative descriptions