Generative AI doesn’t read minds. It can do a lot – but only if you ask the right way. The responses you get from it are only as good as your prompts. Prompting well is a skill worth learning, and you can get better by following some simple rules.
Set the mood
Don’t rush to get a response. Before asking what you want from it, specify how you’d like it framed: thinking about style, tone and perspective.
If you want wit, ask for wit. If you’re gathering information for a Buzzfeed-style listicle, say so. ‘Friendly’ ‘formal’, ‘in the style of my chatty neighbour who’s had a lot of coffee and is obsessed with house plants’ – get creative with how you set the mood.
Give it the backstory
The more context you give - audience, purpose, constraints - the more the AI can act like a collaborator rather than a search engine. If you’re new to this, be clear on that. Tell it who you’re writing for, why, or what to avoid.
Be specific
A vague prompt calls for a vague answer. At best, the AI will choose an angle or direction for you and it will rarely be what you had in mind. Once you’ve set the scene, set to the point: ‘Tell me about plant care.’ will get you a nice overview, but mostly fluff. ‘Explain how to care for my monstera in a dim apartment, using a weekly plan.’ will yield a response you can use right away.
Ask follow-ups
Your first prompt is rarely the best one. Treat AI like a conversation partner. If the answer’s too generic, say ‘make it more detailed,’ or ‘focus on watering frequency.’ Keep digging and refining until you get what you need.
Give real examples
AI learns your style faster when you show it what you like. Paste a snippet of writing you admire. For plant care, you could say, “Write it like this blog post I love, but make it about monsteras.”
Ultimately, prompting isn’t about finding the magic words - but refining your prompts until the response feels right. Start broad, give context, get specific and keep shaping it.
