The correct sequence is Prompt engineering, Retrieval-Augmented Generation (RAG), Fine-tuning.

Prompt engineering is the quickest and lowest effort way to improve responses because you only change the input instructions and context without touching model weights or adding infrastructure. Retrieval-Augmented Generation (RAG) increases implementation complexity because it requires an external retrieval layer and content indexing to provide up-to-date or domain specific context that the model can use to reduce hallucinations. Fine-tuning is the most complex and resource intensive option because it changes model parameters and needs curated training data and ML workflows to produce reliable results.

Retrieval-Augmented Generation (RAG), Prompt engineering, Fine-tuning is incorrect because it places RAG before the simple step of adjusting prompts and it skips the usual progression from quick prompt tweaks to added retrieval and then to model retraining.

Fine-tuning, Retrieval-Augmented Generation (RAG), Prompt engineering is incorrect because it starts with the most costly and permanent step and then moves to simpler options, which reverses the recommended escalation of effort.

Prompt engineering, Fine-tuning, Retrieval-Augmented Generation (RAG) is incorrect because it puts fine-tuning before adding retrieval, but teams commonly try retrieval first to incorporate fresh or proprietary knowledge without the overhead of retraining.

Full AWS Practitioner Certification Question