How to Write Effective Prompts for Work
An effective prompt gives the AI context about your situation, specifies the output format, states who the audience is, and constrains the scope. The key difference between a bad prompt and a good one is specificity: instead of 'write an email about the delay', you describe the situation, the recipient, the tone, and the goal. Iterating on the result is normal and expected.
The quality of your prompt determines the quality of the output. These six steps help you get usable results from ChatGPT, Copilot, or any other AI tool at work.
Why phrasing matters
AI models predict the most likely useful response based on your input. Vague input leads to generic output. Specific input, with context and constraints, leads to output you can actually use. Most people write prompts the way they type a search query: short and keyword-based. That works for Google, but AI models respond better to sentences that describe a situation and a goal.
Step 1: Add context about your situation
Tell the AI who you are, what you are working on, and what you already know. The more relevant context you give, the less the AI has to guess.
Bad: “Write an email about the delay.” Good: “I am a project manager at a software company. Our client expected delivery on June 15 but we need two more weeks. Write a professional email that acknowledges the delay, explains the reason briefly (a dependency issue with a third-party API), and proposes a new delivery date of June 30.”
Step 2: Specify the output format
State what you want to receive: a bulleted list, a table, a formal email, a three-paragraph summary. Without a format instruction, the AI picks one for you.
Bad: “Summarize this report.” Good: “Summarize this report in three bullet points. Each bullet should be one sentence and focus on a key finding.”
Step 3: State your role or audience
Tell the AI who the output is for: a client, a colleague without technical background, or an executive who reads fast. This shapes the tone and vocabulary the AI uses in its response.
Step 4: Give an example of good output
Paste a short example of the kind of result you are looking for. This anchors the AI’s response more precisely than a description alone, particularly for format and tone.
Step 5: Iterate on the result
A first draft is rarely final. Ask the AI to shorten it, change the tone, add a specific section, or rewrite from a different angle. Treat it as a conversation, not a one-shot request. Most useful outputs come after two or three exchanges.
Step 6: Constrain the scope
Tell the AI what to leave out. If you do not want caveats, say so. If the answer should stay under 200 words, specify that. Constraints reduce noise in the output and prevent the AI from adding padding you will have to edit out.
Bad vs good: practical examples
For summarizing a document: vague is “Summarize this.” Better is “Summarize this in five bullet points for a manager who has not read it, focusing on decisions that need to be made.”
For writing an email: vague is “Write a follow-up email.” Better is “Write a follow-up email to a prospect who attended our product demo three days ago. Keep it under 100 words. Do not mention pricing. End with a question about their timeline.”
For data analysis: vague is “What does this data show?” Better is “I have monthly sales data by region for Q1 2025. Which two regions underperformed compared to Q1 2024, and what percentage gap are we looking at?”
If you are using AI inside Power Platform workflows, the same principles apply. See AI Builder in the Power Platform for how prompts work inside automated flows and document processing models.
Our tip: Keep a personal library of prompts that work well for recurring tasks. A simple note or spreadsheet with your best prompts saves time and gives you a starting point instead of starting from scratch each time.
Frequently asked questions
What is a prompt?
A prompt is the instruction you type into an AI tool. It can be a question, a command, or a description of what you want. The AI generates a response based on that input. Unlike a search query, a prompt works better as a full sentence or paragraph that describes the situation, the goal, and the desired format.
Why does phrasing matter so much?
AI models predict the most likely useful response based on your input. Vague input leads to generic output. Specific input, with context and constraints, leads to output you can actually use. Most people write prompts the way they type a search query: short and keyword-based. That works for Google but not for AI models, which respond better to sentences that describe a situation and a goal.
How do I prompt AI for data analysis?
Describe your dataset first: what it contains, how many rows, and what time period it covers. Then state the specific question you want answered. For example: 'I have monthly revenue data by product category for 2024. Identify the three categories with the strongest growth trend and explain what might drive it.' Include the data by pasting it directly if it is small enough, or describe its structure clearly.
Is it bad to iterate multiple times on a prompt?
No. Iterating is normal and often necessary. A first prompt establishes the direction; follow-up prompts refine the result. You can ask the AI to shorten a text, adjust the tone, add a specific section, or remove something you do not need. Treating the AI as a conversation partner rather than a one-shot tool produces better outcomes.