Many academics hope that generative AI can help manage the administrative load that keeps us from meaningful work. But getting useful results from AI assistants requires clear communication – and that means learning to write effective prompts. Just as a new research assistant needs proper guidance to support your work effectively, AI tools need the right context to understand and execute tasks successfully. Writing effective AI prompts is a crucial skill that can help transform AI tools from interesting novelties into genuine productivity partners.

Writing effective AI prompts

The key is moving beyond vague requests to providing structured context. A simple framework can help:

• Role: Give the language model a persona where you tell it something about itself in the context of the response you’re looking for.
• Goal: Explain the outcome you’re looking for.
• Instruct: Tell the language model what you want it to do by giving it examples and steps to follow.

In the video below I describe a framework for prompt writing to get better results from AI assistants like Claude. By providing the right context the AI understands the persona you want it to take on, the goal of the task, and the steps it should follow.

For example, instead of simply asking an AI to “help with student feedback,” you might say:

• Role: You are an experienced lecturer and personal tutor with many years of experience giving supportive and constructive feedback to university students.
• Goal: I need help drafting balanced feedback that acknowledges good work while clearly identifying areas for improvement.
• Instructions: Please 1) review the attached work, 2) highlight 2-3 key strengths, and 3) suggest specific improvements with examples.

Prompts are about context

This structured approach helps ensure the AI understands both the context and your expectations. Just as we wouldn’t throw a stack of papers at a colleague with vague instructions, we shouldn’t expect AI tools to read our minds about what we need.

The “role, goal, instruct” framework is a simple approach to writing effective AI prompts. It provides the context and direction AI needs to support academic tasks. By providing clear context, you can get more reliable output from AI, especially for administrative tasks.

Ready to move beyond basic prompts? Explore the Generative AI for Academics course, where you’ll learn systematic approaches to leveraging AI for your scholarly work.