How to Write AI Prompts: The AliNQuality Prompt Formula

An AI assistant does not read your mind. It responds to what you type. Give it a vague request and it fills the gaps with generic assumptions; give it a clear one and it produces something you can actually use. The first answer is always a draft, and the quality of that draft is set mostly by the quality of your instruction.

The good news is that this is a learnable habit, not a talent. Writing clearer instructions and adjusting them when the result is off is all that prompt engineering really is. You do not need special phrases or technical knowledge. You need a simple structure, which is what the formula below provides.

Consider the difference in practice. Ask an assistant to “write a thank-you email” and you get a generic note that could belong to any business. Ask it to “write a short, warm thank-you email to a customer who just booked their first appointment, mentioning we will send a reminder the day before,” and you get something you can almost send as is. Same tool, same effort to type. The only thing that changed was how much you told it. Every ingredient in the formula is just a prompt to yourself: did I say who, what, and within what limits?

Five ingredients. You do not always need all five, but naming them keeps you from leaving out the part that would have made the difference. Here is each one, with a before-and-after from everyday business work.

Assemble the five parts and you get a prompt that leaves little to chance. Here is a complete example for a common task, replying to a quote request:

Role: You are a helpful assistant for a small landscaping business, writing on my behalf.
Task: Draft a short email replying to a customer who asked for a quote on weekly lawn care.
Context: The customer is Alex, with a medium suburban yard in Nyack. We charge by yard size and visit weekly. I am Sam, the owner. I have not seen the yard yet.
Constraints: Under 120 words, friendly and plain, no exact price since I have not measured the yard, and propose a quick visit to quote accurately.
Output: Give me a subject line and the email body, then note anything I should fill in before sending.

Build your own below. The builder assembles the five parts as you type and gives you a prompt you can copy into any assistant. When you find one that works, save it, or browse ready-made prompts in the prompt library.

A clearer prompt gets a better draft, not a finished product. Before you use any AI output, a person checks four things: facts (names, dates, claims), numbers (prices, totals, calculations), tone (does it land right for the reader), and anything customer-facing that carries your name.

Remember that these tools vary their answers by design, so the same prompt can return different wording on different runs and in different tools. That is normal, and it is the reason review is part of the workflow rather than an afterthought. You stay responsible for what is true, what adds up, and what represents your business.