Why Most People Get Mediocre Results from AI
Most people treat AI tools like a search engine — they type a vague question and hope for the best. When they get a generic answer, they assume AI isn't that useful. The problem isn't the AI. The problem is the prompt.
Prompt engineering is simply the skill of giving AI tools clear, specific instructions so they can do their best work. You don't need to be a programmer to do this well. You just need to understand a few principles.
The 4 Core Elements of a Great Prompt
Role
"Write a marketing email."
"You are a senior copywriter who specializes in B2B SaaS. Write a marketing email..."
Why it works: Give the AI a role that matches the expertise you need. This calibrates its tone, vocabulary, and approach.
Context
"Summarize this report for me."
"I need to present this quarterly financial report to non-financial stakeholders at a board meeting. Summarize the key points in plain language, highlighting the 3 most important takeaways."
Why it works: The more context you give, the more tailored the result. Tell the AI who the audience is, what the goal is, and any constraints.
Format
"Give me ideas for our content strategy."
"Give me 5 ideas for our content strategy. Format each idea as: Title, 2-sentence description, target audience, and expected outcome."
Why it works: Be specific about the output format. Want a table? Bullet points? A numbered list with sub-items? Say so explicitly.
Constraints
"Write a bio for me."
"Write a professional bio for a LinkedIn profile. Keep it under 150 words. Use third person. Focus on leadership and results rather than job titles. Avoid buzzwords like "passionate" or "results-driven"."
Why it works: Tell the AI what NOT to do, as well as what to do. Constraints lead to sharper, more usable outputs.
Advanced Technique: Chain of Thought
For complex tasks, tell the AI to think step by step before giving you an answer. This dramatically improves accuracy on anything that requires reasoning.
"Before giving me your recommendation, think through this step by step: (1) What are the key factors to consider? (2) What are the tradeoffs of each option? (3) Then give me your recommendation with a clear rationale."
Real-World Templates You Can Use Today
Drafting an email
You are a professional business writer. Draft a [type] email to [recipient] about [topic]. The tone should be [formal/friendly/direct]. Keep it under [length]. The goal is to [outcome]. Do not use [any phrases to avoid].
Summarizing a document
Summarize the following [document type] in [X] bullet points. Write for a [audience] who [context about their knowledge level]. Focus on the most actionable insights. [Paste document]
Brainstorming ideas
Generate [number] ideas for [topic]. For each idea, include: a one-sentence description, who it's best for, and one potential downside. I'm trying to [goal]. My main constraint is [constraint].
Reviewing your writing
Review the following [document type] and provide specific feedback on: (1) clarity, (2) tone for the intended audience of [audience], (3) any missing information, and (4) suggested improvements. Then provide an edited version. [Paste writing]
Key Takeaways
- Give the AI a role that matches the expertise you need
- Provide context: who is the audience, what is the goal, what are the constraints
- Specify the format you want the output in
- Tell the AI what to avoid, not just what to do
- For complex tasks, ask it to think step by step
- Treat it like briefing a capable colleague, not typing into a search box
Get More Guides Like This
Subscribe to our free weekly newsletter for practical AI tips and news.
Subscribe free