An airline wants a text-based conversational assistant (using LLMs and a knowledge base) to answer customers about schedules, bookings, and payments with minimal development effort. Which approach requires the least development work?

Building a RAG agent with Amazon Bedrock requires the least development effort for this use case. Bedrock provides managed access to powerful LLMs, and RAG agents integrate these LLMs with a knowledge base (like airline FAQs or booking data) to provide accurate, context-aware answers without extensive model training. This approach leverages pre-trained models and a structured framework for retrieval and generation. Training models with SageMaker Autopilot or fine-tuning with SageMaker JumpStart involves more data preparation, training, and evaluation, increasing development time. Developing a custom Python application with Amazon Q Developer would also entail significant custom coding and integration work.