If you’ve ever wondered how to prompt AI in a way that gets you smarter, sharper, and more reliable results, you’re not alone. Prompts are the fuel that power AI responses, and with the right approach, you can take outputs from “meh” to mind-blowing. Here are 10 practical tips for AI prompting (plus a bonus) that actually work.
10 Practical Tips on How to Prompt AI Effectively
1. Assign a Role or Persona
Why it works: Asking 'Write a sonnet about love and time' is not the same as asking 'You are William Shakespeare. Write a sonnet about love and time'. AI thrives on clarity. When you give it a role, you instantly shape the perspective, tone, and style of the response. Instead of staying generic, it adapts to the mindset of the role. It’s like hiring a specialist for free.
2. Describe the Context Concisely
Why it works: Context is everything in effective prompting. Without it, AI often defaults to broad or irrelevant answers. Providing specific details or background information makes a massive difference, ensures the model uses relevant knowledge and avoids generic or ambiguous responses.
3. Define the Task Clearly Starting with a Verb
Why it works: "Summarise", "list", "classify", "compare", "rewrite", "translate", "write"... Action verbs eliminate ambiguity and explicitly tells the LLM what action to perform. Vague prompts lead to vague answers, but specific tasks produce structured results. Think of it like giving directions: “Turn left” is clearer than “Go somewhere.”
4. Set Constraints and Output Lengths
Why it works: AI can easily over-explain or ramble without boundaries unnecessarily. By setting clear constraints (e.g. word limits, number of bullet points, or style guidelines) helps manage the LLM's response size. This reduces drastically response time, computational resources, and costs. It also prevents the model from generating excessively long or irrelevant text.
5. Define the Output Format or Structure
Why it works: Specifying the desired output format (e.g., JSON, XML, bullet points, numbered steps, table) ensures that the LLM delivers information in a structured and usable manner. Instead of reformatting later, you get neatly packaged results tailored to your workflow.
6. Use One-Shot or Few-Shot Examples
Why it works: AI learns patterns from examples. Giving one (one-shot) or a few (few-shot) samples guides its tone, structure, and depth. It’s like saying, “Do it like this.” Instead of leaving interpretation open, you anchor the AI to a clear style, making the results more predictable and consistent.
7. Use Natural Language and Conversational Style
Why it works: You don’t need to “speak robot” to get good results. AI responds best when you talk to it naturally, as if you were messaging a friend or colleague. This approach leverages the model's training on vast amounts of natural text. Overcomplicating prompts often makes results worse. Plain English leads to clearer communication, fewer misunderstandings, and more human-like outputs.
8. Consider Audience and Knowledge Level
Why it works: Would you talk to your boss the way you talk to your kid? A beginner, a student, and a professional each need different explanations. When you specify the audience, AI adjusts complexity, tone, and detail. This makes the content more relatable and easy to understand, whether you’re targeting casual readers or industry experts.
9. Save Prompt Templates That Work
Why it works: Great prompts are like recipes - you don’t want to reinvent them every time. By saving templates, you build a personal library of proven approaches. This makes your workflow faster, more consistent, and more reliable. Over time, you’ll have a go-to toolkit that gives you strong results on demand.
My Prompt Templates Toolkit:
"Summarise any article in 3 bullet points. [Insert article text here]"
"Generate 5 catchy blog titles about [Insert topic here]"
"Explain [Insert concept here] in simple terms for a 12-year-old"
10. Follow the RAFT Prompt Structure
Why it works: RAFT = Role, Audience, Format, Task ("Rodrigo Ate Four Tacos 🌮" is easier to remember). It’s a simple yet powerful framework that makes prompting easy to remember and highly effective. While not all prompts require every component, combining at least a clear task with one or more supporting elements makes instructions easier for LLMs to follow. Think of it as a safety net for consistently strong results.
Bonus Tip: Use Figures of Speech
Why it works: By using metaphors, similes, irony, or rhetorical questions in your prompts, you nudge it to produce more engaging, human-like, and memorable responses. Figures of speech spark imagination and help AI break free from overly dry or technical phrasing, making outputs fun to read and easy to understand.
Try this one yourself =)
“Very briefly explain step by step how to cook lentils using British street slang used by young people in London”
Final Thoughts
Mastering how to prompt AI isn’t about being technical, it’s about being clear, specific, and structured. The more you practise with these tips for AI prompting, the better your results will get. Don’t just read these tips, put them to work!
👉 Want more ready-to-use prompt templates and tips? Subscribe to our newsletter and keep sharpening your AI superpowers.