Prompting Pitfalls

Prompting sounds easy - just ask a question, right?

But most people make common mistakes without realizing it.

Here are some of the most common ways prompting goes wrong:

Vague or Ambiguous Prompts like “Help me write an email” leave too much room for interpretation. The AI doesn’t know what the email is for or what it should sound like.

Missing Context When you leave out background, audience, or purpose, the AI will fill in the blanks - just like a person would.

Unclear Vision If you’re not sure what you want, the AI won’t be either. Without a goal or direction, you’ll get unfocused output or hallucinations.

Unclear Expectations Forgetting to include tone, length, format, or point of view leads to generic responses.

Command or Q&A Centric Thinking Many users treat prompting like asking Siri: one question, one answer. But LLMs are built for deeper, step-by-step exploration. If you’re only giving commands or asking quick-fire questions, you’re not tapping into their full potential.

The fix? Be thoughtful, not just quick. Add structure, explain your needs, and treat prompting like the beginning of a conversation- not the end of one.

There are some well-known prompting techniques that can help.