Provide context, specify constraints, and show examples.
The quality of an AI's response is directly linked to the quality of the prompt. Learning basic prompting techniques—providing relevant context, defining clear output constraints, and showing examples (few-shot prompting)—helps ensure relevant, structured results.
Prompting is a practical communication skill. To get the best results from language models, avoid simple, single-sentence queries. Instead, structure your prompt by defining the AI's role, providing context, listing strict output constraints (such as length or format), and showing examples of your expected output style.
Prompt the model to extract names and dates from a text block, specifying that the output must be formatted as a clean JSON array with no conversational intro.
Provide a model with 2-3 of your past newsletters as examples, prompting it to draft a new announcement matching that specific tone and structure.
Few-shot prompting involves including 1-2 examples of your expected input and output style within the prompt, helping the model align with your formatting goals.