Prompting vs. Prompt Engineering
Prompting is the act of interacting with a generative AI system through natural language input. Any time you write a query to an LLM, you’re prompting.
Prompt Engineering is the iterative process of developing and optimizing prompts to achieve reliable, high-quality outputs for specific use cases. It involves systematic experimentation, evaluation, and refinement.
The distinction matters because prompting is accessible to anyone, while prompt engineering is a disciplined practice. Casual prompting often suffices for simple tasks. Complex, high-stakes, or production applications typically require engineering, systematic testing against evaluation sets, version control, performance benchmarking.
A prompting technique is a specific method for structuring prompts (Chain-of-Thought, few-shot, role prompting). A prompt engineering technique is a method for developing and optimizing prompts (automated search, A/B testing, meta-prompting).
The field conflates these terms constantly, which creates confusion about what’s actually being discussed.
Related: 05-atom—prompt-component-taxonomy, 05-atom—ai-vs-human-prompt-optimization