How to Write Prompts That Get You What You Want from AI
By Alpha Covenant Team · 2026-08-30
What's Actually Happening When You Type to an AI
When you send a message to an AI chatbot, the system doesn't "think" the way a person does. It reads every word you typed and generates a response by predicting what helpful text should follow — based on patterns learned from enormous amounts of written material. That means the AI has no idea what's in your head. It only works with what you actually wrote. A vague question gets a vague answer. A clear, specific question gives the AI the raw material it needs to produce something genuinely useful. The technical word for your message is a "prompt," but all that really means is: the instructions you give.
What AI Chatbots Are Genuinely Good At
When prompted well, AI chatbots are remarkably useful for drafting, explaining, brainstorming, and organizing. They don't get tired, they don't judge you, and they can shift tone or approach the moment you ask.
A concrete everyday example: Say you need to write a complaint email to your landlord about a broken heater, but you're not sure how to strike the right tone — firm but not aggressive. You could type: "Write a polite but firm email to my landlord asking him to fix the broken heater in my apartment. The heater has been out for two weeks. Keep it professional and under 150 words." That prompt gives the AI a role (write an email), a recipient (landlord), a specific problem (broken heater, two weeks), a tone (polite but firm), and a length limit. The result will almost certainly be usable — probably better than staring at a blank screen for 20 minutes.
What to Watch Out For
AI makes things up — confidently. This is the single most important risk. If you ask a vague question, the AI may fill gaps with invented details that sound completely real. Dates, names, statistics, quotes — all can be fabricated. Never copy AI output on factual matters without checking an independent source.
Garbage in, garbage out. A muddled prompt produces a muddled response. If you ask something like "help me with my thing," the AI will guess at what you mean, and it will often guess wrong.
The AI doesn't know your situation. It doesn't know your landlord is actually pretty reasonable, or that you need to keep things informal, or that there's legal history involved. It only knows what you tell it. If you leave out important context, you'll get a generic response that might not fit your real circumstances at all.
It can agree with you too easily. AI chatbots tend to be agreeable. If you push back on an answer — even if the AI was originally correct — it may cave and tell you what you want to hear. Treat AI opinions as a starting point, not a verdict.
Privacy matters. Don't paste sensitive information — Social Security numbers, passwords, medical records, private details about other people — into any AI chatbot unless you've carefully read that service's privacy policy. Assume what you type could be stored or reviewed.
How to Use It Well — Step by Step
1. State what you actually want, not just a topic. Weak: "Email." Better: "Write an email to my landlord about a broken heater." Best: "Write a polite but firm email to my landlord asking him to fix the broken heater. It's been broken two weeks. Keep it under 150 words."
2. Give it a role if it helps. Starting with "Act as a friendly financial advisor" or "Explain this as if I'm twelve years old" genuinely shifts how the AI responds. You're not flattering it — you're calibrating its output.
3. Add the context it can't guess. Who is the audience? What's the goal? What tone do you want? What should it avoid? The more you tell it, the less it has to invent.
4. Set a format or length. "Give me a bullet-point list" or "Keep it under 100 words" or "Write three short options I can choose from" — these constraints make the output far more usable.
5. Check and edit the result. Never treat AI output as finished work. Read it. Ask yourself if it's accurate. Change what doesn't fit. The AI gave you a draft — you make it yours.
6. Ask it to try again if the first answer misses. You can say: "That's too formal — try again, more casual" or "You missed the point — I need X, not Y." Treat it like a back-and-forth, not a one-shot machine.
Prompt of the Brief
Here is one template you can copy, fill in, and use immediately for almost any writing task:
Act as a [role, e.g. "professional editor" / "friendly advisor" / "plain-English explainer"].
I need you to [task, e.g. "write" / "summarize" / "rewrite" / "explain"].
Here is the context: [describe your situation, who it's for, what's already happened].
Tone: [e.g. "warm but direct" / "formal" / "casual and friendly"].
Format: [e.g. "bullet points" / "short paragraphs" / "under 200 words" / "three options"].
Please avoid: [anything you don't want — jargon, length, a specific approach].
Why it works: It removes every major source of guessing. The AI knows its role, the task, your context, the tone you need, and what to skip. Each blank you fill in is one less thing the AI has to invent. To adapt it: You don't need every section every time. For a simple request, role + task + tone is often enough. Add more fields when the stakes are higher — a job application letter deserves more detail than a grocery list organizer.
Want this handled for you?
Alpha Covenant runs your entire demand engine — finding, pitching, and closing your next high-value client, autonomously and at cost, with full receipts.
👉 Run your free visibility audit and see exactly where your buyers are looking for you — or activate instantly.
This article was produced with the assistance of AI and reviewed by our team.