Effective AI prompts are clear, specific, and provide context. Including additional details such as tone (formal, conversational, or technical), style, and audience improves the quality of your prompts. And the quality of your prompts directly impacts the quality of the AI-generated output. Ambiguous or poorly constructed prompts can result in irrelevant or incomplete outputs.
This blog provides six quick tips for writing effective prompts.
1. Be Specific:
Instead of vague requests, clearly define what you want AI to generate. Avoid ambiguous words or phrases that could lead to unintended results. Also, avoid true/false type questions to which AI may give binary answers.
You should write your prompts in a way that mimics everyday speech. Ask for specific actions or targeted answers rather than general or open-ended requests.
Provide an image of a small desk organizer that holds a cell phone, features a charging station, and includes storage for office stationery. Example: The desk organizer should have a designated slot for a smartphone that includes a wireless charging pad, individual compartments for pens and pencils, sticky notes, as well as a small drawer for paper clips.
Vague Prompt: “How does document collaboration feature work?”
Clear Prompt: “How does the document collaboration feature in Microsoft Word enable multiple users to edit and comment on the same file in real time?”
2. Provide Context:
Providing background information or context can help AI process your request better.
By providing context, AI tailors its response to better suit your needs, making the outcome more aligned with your objectives.
Here are two examples to emphasize the importance of clarity and precision when crafting AI prompts.
Example 1:
Instead of “Request a meeting with [contact name],” you might say,
“As a project manager, request a bi-weekly meeting with [team lead] to review the project status.”
Example 2:
Prompt without context: “Summarize the meeting notes.”
Prompt with context: “As a marketing analyst preparing a report for the executive team, summarize the key action items and decisions from last week’s strategy meeting, highlighting any marketing-related tasks.”
The second prompt shows how a vague request like “Summarize the meeting notes” can be transformed into a clear directive by providing context and including relevant details (such as your role and specific objectives).
3. Use Constraints to Narrow the Scope:
If you ask general questions or make open-ended requests, AI may provide broad and irrelevant responses. Specifying constraints within your prompt, such as setting, tone, style, format (including file type, word limits, etc.) helps in narrowing down the response. If you are writing for a specific audience, tailor your prompt accordingly. This approach can help AI generate output that is closer to your desired result.
Consider the following example on how adding constraints can turn a vague image generation request into a clear directive, improving results.
Objective:
Create engaging and festive marketing content for a holiday campaign for a pet company. Feature a golden retriever in a Christmas setting, complete with a red Santa hat, to evoke warmth and holiday cheer, attracting pet lovers and enhancing brand appeal during the seasonal promotion.
Initial Prompt: “Create a holiday portrait of a dog.” 
This initial prompt lacks specificity, leaving AI with a broad interpretation. The generated response might include a golden retriever but may fail to incorporate other key elements associated with the holiday theme, such as festive accessories or colors. The output risks being generic and missing the intended spirit of the request.
Refined Prompt: “Create a high-quality, festive Christmas marketing image featuring a friendly golden retriever sitting in a cozy, decorated living room. The dog is wearing a bright red Santa hat and is surrounded by wrapped presents and twinkling holiday lights, conveying warmth and holiday cheer. The overall scene should be inviting and joyful, designed to appeal to pet lovers and enhance brand appeal during holiday campaign.” 
This refined prompt provided AI with the following details and constraints:
-Purpose: Specifies the image is for marketing and should be high-quality.
-Setting: Includes a cozy, decorated living room to create a festive atmosphere.
-Details: Adds holiday props like presents and lights for visual appeal.
-Tone: Requests an inviting and joyful mood to evoke warmth and cheer.
-Target audience: Explicitly mentions pet lovers and brand appeal.
By adding these details, the prompt becomes more focused and guides AI to include festive elements and create a specific, visually appealing image that matches your expectations. The constraints ensure that the output embodies the desired theme and aesthetic, leaving little room for ambiguity.
4. Revise and Refine:
Keep experimenting if you’re not getting the desired results, and refine your prompts based on AI's responses. This is an iterative process. You can rephrase your prompt and provide more context to get better results.
Below is an example of refining your prompts to achieve the desired results.
Objective:
The objective is to create a basic HTML table without outer borders and only light grey inner borders and no other style specifications.
Not So Great Prompt:
“Create a table with borders inside the cells.”
This prompt is vague and does not specify the type, size, or style of the borders. It also does not mention the absence of outer borders or the lack of additional styling. AI might generate a table but is highly unlikely to produce the desired results.
Somewhat Improved Prompt:
“Generate an HTML table with 1pt solid borders inside the cells, but no outer border.”
This revised prompt is clearer and introduces border size and type, but it still lacks precision. It fails to specify the color of the borders and does not explicitly request that no additional styling be applied. AI’s output might get closer to the goal but could still include unwanted features.
Optimal Prompt:
“Generate RTE HTML code for a table where only the inner cell borders are visible, using 1pt solid #7E7E7E lines. Ensure that there are no outer table borders and do not apply any additional styling beyond these specifications.”
This prompt is specific and unambiguous, and AI can generate the HTML code snippet for the table based on this clear and concise instruction.
5. Provide Examples as Guidelines:
Providing examples in your prompt can help guide AI towards the desired output. This is useful for tasks like generating text in a specific style, writing code, or explaining concepts. You could also add links to helpful references as shown below.
“Summarize the main points of this article: [link to article]”. Adding links to references in your prompt enables AI to access specific sources, resulting in more accurate and contextually relevant responses.
Without Example Prompt:
“Design an image of a desk organizer that can hold a cell phone and office stationery.” 
With Example Effective Prompt:
“Provide an image of a small desk organizer that holds a cell phone, features a charging station, and includes storage for office stationery. Example: The desk organizer should have a designated slot for a smartphone that includes a wireless charging pad, individual compartments for pens and pencils, sticky notes, as well as a small drawer for paper clips.”
Give Feedback:
After AI generates an output, if the response isn’t exactly what you were looking for, provide feedback to help it learn and improve. AI learns from your feedback.
You could say, “You covered A in detail, but I needed more information on B. Could you clarify more details about xyz?"
With consistent practice, you can create prompts that will guide AI to generate output in your preferred writing style. The more constraints, context, and direction you provide with your feedback, the better AI can comprehend your requirements to generate the desired output.
About the Author
Follow on Linkedin More Content by Surya Nair
