Better prompts aren't magic incantations but rather clearer expressions of what you actually need: naming your context and constraints, specifying output format, identifying what you want to avoid, and asking follow-up questions when answers miss. Good prompting is simply good communication with a tool.
You've probably noticed that sometimes ChatGPT gives you brilliant help, and sometimes it gives you generic nonsense. The difference usually isn't the AI—it's how you asked the question. This is called prompt engineering, and it's just the art of asking better questions.
Think of it like asking a friend for directions. If you say "How do I get to campus?" they might give you basic directions. But if you say "I'm at the library, I need to get to the science building, and I have 10 minutes, what's the fastest route?", you get way more useful information. The AI works the same way.
Good prompts have four ingredients: context (what's the situation?), specificity (exactly what do you need?), format (how should the answer look?), and constraints (any limitations?). For a college student, this might look like: "I have a 500-word essay due on climate change for Environmental Science 101. My professor emphasizes peer-reviewed sources and real-world policy applications. Can you help me outline an essay that covers both?" versus just "write about climate change."
The context helps the AI understand your actual constraints. The specificity prevents generic answers. The format request means you get something actually usable, and constraints keep the AI from wasting words on irrelevant information.
Here's where it gets powerful: You can embed your professor's teaching style into your prompts. If you know they care about applications over theory, you could say "Help me study for my chemistry exam the way Professor Lee teaches it—with real-world lab applications, not just theory." The AI has a target now.
Another game-changer is telling the AI to think step-by-step. If you're struggling with a complex problem, ask: "Walk me through how to approach this step-by-step" instead of just asking for the answer. This actually makes the AI more helpful because it's explaining reasoning, not just delivering information.
Common mistakes: Being too vague, not giving enough context, asking the AI to do something it's not designed for (like provide current news from 2024 if the AI's training data is from 2023), or not iterating. If the first answer isn't helpful, ask follow-up questions. Refine. That's engineering—making something work better through repeated adjustment.
Try this: Take an assignment you're actually working on right now. Write two prompts for it—one vague (like you'd normally ask) and one detailed (with context, format, and constraints). Compare the results. You'll immediately see why prompt engineering matters.
Peri can explain this concept, give practical examples, help you decide whether it applies to your situation, or recommend a journey if appropriate.
Explore related journeys or tell Peri what you're working through.