Why Prompts Matter
The quality of AI output depends entirely on the quality of your input. A vague prompt produces vague results. A well-structured prompt produces focused, actionable content that respects journalistic standards.
Prompt engineering is the skill of crafting instructions that guide AI models toward the specific output you need — whether that is a story angle, a data analysis, or a fact-check workflow.
The CRAFT Framework
Five elements of an effective journalism prompt.
Context
Provide background information. What is the story about? What do you already know? Give the AI the context it needs to respond accurately.
Role
Define who the AI should act as. An investigative reporter? A data analyst? A copy editor? The role shapes the perspective and expertise of the response.
Action
Specify the task clearly. Summarize, analyze, compare, draft, fact-check — use precise verbs that leave no ambiguity about what you need.
Format
Define the output structure. Bullet points, a 200-word summary, a table, an inverted pyramid article — tell the AI exactly how to organize the response.
Tone
Set the voice. Objective and formal for hard news, conversational for features, analytical for data stories. Tone ensures the output matches your editorial style.
Example Prompts
See the CRAFT framework in action with these journalism-specific examples.
Interview Preparation
Role: Act as an experienced political reporter.
Action: Generate 10 tough but fair interview questions.
Format: Numbered list with follow-up suggestions.
Tone: Professional, probing, non-partisan.
Data Analysis
Role: Act as a data journalist.
Action: Identify the three most significant trends and explain their news value.
Format: Brief analysis with bullet points for each trend.
Tone: Analytical, precise, evidence-based.
Journalaism Learning Program
A structured, self-paced learning program that takes you from AI novice to confident practitioner. Includes assessments, exercises, and certification.