Prompting Basics: How to Talk to AI
The single biggest factor in getting good results from AI isn't which model you use — it's how you ask. Here are the fundamentals that will immediately improve every AI interaction you have.
1. Be Specific, Not Vague
The most common mistake people make with AI is being too vague. AI models are literal — they'll do exactly what you ask, so the more specific your request, the better your result.
Vague Prompt
"Write about dogs"
Specific Prompt
"Write a 200-word blog introduction about why golden retrievers make great family pets for households with young children"
Notice the difference? The specific prompt tells the AI the format (blog introduction), the length (200 words), the topic (golden retrievers as family pets), and the audience (families with young children). The AI has everything it needs to give you exactly what you want.
2. Give Context
AI doesn't know who you are, what you're working on, or why you're asking. The more context you provide, the more relevant the response will be.
Example with Context
"I'm a small business owner running a bakery in Manchester. I need to write a polite but firm email to a supplier who has delivered the wrong flour for the third time this month. The tone should be professional but make clear this is the last chance before we switch suppliers."
Context includes: who you are, what you're doing, who the audience is, the tone you want, and any constraints. Think of it like briefing a new colleague — the more they know, the better they can help.
3. Specify the Format
Don't just ask for information — tell the AI how you want it structured. This saves you from having to reformat the output.
"Give me this as a bullet-point list"
"Present this as a table with columns for Name, Price, and Pros/Cons"
"Write this as a step-by-step numbered guide"
"Keep the answer under 3 paragraphs"
4. Tell the AI Who to Be
One of the most powerful prompting techniques is assigning the AI a role. This frames its entire response through the lens of that expertise.
Role Examples
- "You are an experienced primary school teacher..."
- "Act as a senior financial adviser..."
- "You are a food critic writing for a newspaper..."
- "Respond as a patient, encouraging fitness coach..."
Why This Works
When you tell the AI to "be" a teacher, it adjusts its vocabulary, examples, and explanations to match how a teacher would communicate. The same question answered by a "university professor" vs. a "primary school teacher" will be dramatically different in complexity.
5. Iterate and Refine
You don't have to get it right in one go. Treat AI like a conversation, not a search engine. The first response is a starting point — then you refine.
"Write me a cover letter for this job posting" → too generic
"Make it more specific to my 5 years of marketing experience" → better
"Tone down the enthusiasm and make it sound more confident and understated" → great
"Shorten the second paragraph and add a specific achievement with a number" → perfect
6. Say What NOT to Do
Constraints are just as important as instructions. Telling the AI what to avoid is often the fastest way to fix a common problem.
- "Don't use bullet points" — forces narrative prose
- "Avoid jargon — explain this for a complete beginner" — controls complexity
- "Don't start with 'Sure!' or 'Great question!'" — removes filler
- "Do not make up statistics — say 'I'm not sure' if you don't know" — reduces hallucination