Prompt Engineering
Clear instructions help — but prompting is only one part of building a reliable AI workflow.
What is prompt engineering?
Prompt engineering is the practice of structuring inputs so an AI system has the context, task, constraints, and output requirements it needs.
Good prompts often answer five questions:
- Context — What should the system know?
- Task — What should it do?
- Constraints — What rules or boundaries apply?
- Output — What form should the result take?
- Verification — What evidence, tests, or checks are required?
Examples help when the pattern matters
Few-shot prompting supplies examples of the input/output behavior you want. There is no universal “optimal” number of examples: useful count and ordering depend on the model, task, context budget, and diversity of cases.
Break complex tasks into checkable stages
For multi-step work, ask for outputs you can inspect:
1. State the assumptions.
2. Perform the calculation.
3. Show the intermediate values needed to verify it.
4. Give the final answer.
5. Flag any uncertainty.
This is more useful than treating a long internal reasoning trace as proof of correctness.
References
Reynolds, L., & McDonell, K. (2021). Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm. CHI EA 2021.
Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., et al. (2022). Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. Advances in Neural Information Processing Systems (NeurIPS).
Zhou, Y., Muresanu, A. I., Han, Z., Paster, K., Pitis, S., et al. (2022). Large Language Models Are Human-Level Prompt Engineers. ICLR 2023.
Citation Note: Some citations are open access (arXiv or DOI, where linked). Others are books or journal articles that may sit behind a publisher paywall. Use the linked DOI or arXiv when available. If you notice any citation errors, please let us know.