Question

A logistics company wants to predict, for each of the 50,000 shipments processed daily, the probability that a package will be delivered late based on historical structured data such as route, weather, carrier, and package weight. A director suggests using a generative AI large language model for this task. As the AI Transformation Leader, what is the most appropriate recommendation?

Accepted Answer

Use a traditional machine learning classification or regression model, because the task is a structured numerical prediction where generative AI adds little value

Answer

Use a generative AI model with detailed prompt engineering to produce the probability scores for each shipment

Answer

Use a generative AI model grounded with retrieval-augmented generation over the shipment database

Answer

Use a fine-tuned generative AI model trained on the historical shipment records