A team wants a foundation model to follow a specific output format and tone. What's the most cost-effective FIRST approach before considering fine-tuning?
A.Train a new model from scratch
Scratch training is enormous cost — never the first move.
B.Prompt engineering with instructions, examples (few-shot), and output format cues
Correct. Cheapest, fastest lever — try it first.
C.Increase the model's context window
Window size doesn't teach format or tone.
D.Move to a larger model
Bigger models may follow instructions better but cost more — try prompts first.
Why this answer
The standard escalation ladder: prompt engineering → RAG (for missing knowledge) → fine-tuning (for style/format/behavior RAG can't fix). Prompt engineering is free, instant, and solves most formatting/tone needs. Fine-tuning costs training time, data prep, and hosting premium.
Correct answer: B · Last reviewed 2026-11-02
Official AWS references
Practice this interactively
Answer without seeing the solution first in the Generative AI Concepts practice session.
Practice Generative AI Concepts