A Practical Guide to Prompting
Do you ever feel like you’re not getting the best out of AI chatbots? You’re not alone. While tools like ChatGPT, Gemini and Claude are incredibly powerful, getting good results requires understanding a few key principles. Here are some practical tips to help you have more productive AI conversations. I’ve written about many of these at one time or another, but it’s useful to gather them in one place. Note that while the various chatbots act a little differently, these guidelines should be applicable to any of the major AI chatbots.
A little time spent working with a few guidelines will make your interactions with AI more efficient and (more importantly) more effective. Remember, these are guidelines, not laws, so feel free to experiment to figure out what works best for YOU. (Note: There are some overlaps below. That’s intentional.)
Start Simple, Then Refine
Writing complex, highly-engineered prompts is typically not necessary. So, don't feel pressured to write complex prompts. Instead:
Begin with a straightforward question
Read the AI's response carefully
Ask follow-up questions to get more detail or clarification
Guide the conversation toward what you need
If response quality seems to be degrading, ask for a summary of the conversation, then start a new chat session.
This approach, called iterative prompting, often works better than trying to craft the perfect prompt upfront.
Be Clear About Your Goals
Before asking AI for help, consider:
What specific outcome do you want?
Who is the intended audience?
What level of detail do you need?
How will you use the information?
What output format do you want?
Including these details helps AI provide more relevant and useful responses.
Provide Helpful Context
AI doesn't know your specific situation unless you explain it. Share relevant details like:
Your background knowledge of the topic
Any constraints or requirements
Examples of what you're looking for
What you've already tried
Your target audience
Any potentially helpful data or documents
What role you want AI to play (colleague, mentor, expert, etc.)
For instance, instead of asking "How do I learn programming?" try "Act as an expert coding tutor. I'm a marketing professional interested in learning Python for data analysis. I have no coding experience. What's a good way to start?" Generally, it’s better to provide too much context than too little, but this isn’t always the case.
Use the Right Approach for Your Task
Different tasks need different approaches:
For creative brainstorming: Keep prompts open-ended and invite creativity.
For specific information: Be precise in your questions.
For analysis: Ask for step-by-step reasoning.
For practical advice: Request concrete examples and the rationale behind the advice
For summaries: Define the audience, purpose, length and level of detail.
Meta-Prompting is Your Friend for Complex Tasks
For complex tasks, or even when you just don’t know how to get started, ask the chatbot to help you. This is called “meta-prompting” and it’s often highly effective. I use this frequently for deep research tasks. Remember, though, this isn’t magic. You still need to review and refine AI’s prompt.
Describe your task, being specific about your goal/objective.
Provide detailed context.
Ask AI to write a prompt to carry out your task and accomplish your goal.
Refine the prompt as necessary either on your own or with AI’s help.
Use the prompt as your starting point.
Don't Just Accept Everything
AI can be confidently wrong. When using AI:
Verify factual claims, especially for important decisions.
Cross-check technical information.
Be especially careful with dates and statistics.
Ask AI to provide references with links, but remember to check that the references exist.
Use AI's suggestions as a starting point, not the final word.
Learn From the Conversation
Pay attention to what works and what doesn't. For example, if you notice AI often going in unhelpful directions, you may need to provide more details about your goal and more context. During your AI chat sessions:
Notice which types of questions get better responses.
Pay attention when AI asks for additional information or takes a wrong path.
Keep track of effective prompting patterns.
Learn from misunderstandings and unclear responses.
Adjust your approach based on results.
Consider the Nature of the Task
When working on documents:
Share relevant background materials.
Be specific about the type of feedback you want.
Provide AI with a role (co-author, journal editor, reviewer, dean, etc.).
Ask for explanations when you don't understand suggestions.
Request examples to clarify points.
For problem-solving:
When possible, use reasoning models (e.g., ChatGPT o3, Gemini 2.5, Claude 3.7). These models are better and step-by-step reasoning through complex tasks and problems.
Break complex problems into smaller parts.
Ask AI to explain its reasoning.
Request alternative approaches.
Use AI to check your work.
For brainstorming and refining ideas:
Clearly define your goal.
Provide relevant context and constraints.
Start with broad, open-ended questions.
Ask AI to approach the task or problem from different perspectives.
When zeroing in on the best ideas, ask AI to explain its reasoning.
Iterate, iterate, iterate.
Common Pitfalls to Avoid
There are a few common pitfalls you’ll want to avoid:
Don't expect perfection on the first try.
Avoid overly complex or vague prompts.
Don't assume AI understands context it hasn't been given.
Be cautious about technical or legal advice.
Remember that even the best models hallucinate.
When Things Go Wrong
If you're not getting useful responses:
Rephrase your question.
Provide more context.
Break down complex requests into simpler ones.
Ask AI what additional information would be helpful.
Start a fresh conversation if you're going in circles.
Analyze the chat transcript to see where things went wrong. This can help you learn to better guide AI.
Sidebar: AI Assignment Repository Survey
If you haven’t already done so, I would appreciate it if you would complete this short survey gauging interest in a free, open-source repository of learning activities that discourage inappropriate AI use and/or leverage AI to enhance student learning.
Wow. That’s a lot to remember. Let’s distill this advice down to a handful of simple guidelines in a too long; didn’t read (TL;DR) version.
Getting Better Results from AI: The TL;DR Guide
Want better results from AI chatbots? Here are the key tips:
Keep it simple: Start with basic questions, then refine through follow-ups
Be specific: Tell AI your goal, audience, and how you'll use the information
Give context: Share your background and requirements - AI can't read your mind
Verify everything: AI can be confidently wrong, so fact-check important claims
Match your approach to your task: Use open prompts for creativity, specific ones for facts
When stuck: Rephrase, add context, or start fresh
Learn patterns: Notice what works and adjust your approach accordingly
Let AI help you: Use meta-prompting for complex tasks
Remember, AI is a tool to enhance your thinking, not replace it. The goal is to use AI to help you work more effectively, not to do everything for you. With practice, you'll develop an intuitive sense of how to get the most out of these powerful tools.
Notes:
Much of this was derived from https://simonwillison.net/2025/May/25/claude-4-system-prompt. Lex.page helped quite a bit with this one.
Want to continue this conversation? I'd love to hear your thoughts on how you're using AI to develop critical thinking skills in your courses. Drop me a line at Craig@AIGoesToCollege.com. Be sure to check out the AI Goes to College podcast, which I co-host with Dr. Robert E. Crossler. It's available at https://www.aigoestocollege.com/follow.
Looking for practical guidance on AI in higher education? I offer engaging workshops and talks—both remotely and in person—on using AI to enhance learning while preserving academic integrity. Email me to discuss bringing these insights to your institution, or feel free to share my contact information with your professional development team.
Well written. I just wrote about meta cognition in my blog post The Skill Everyone Overlooks in AI also