6 Basic Building Blocks for Marketers to Write Effective AI Prompts
AI models have quickly become powerful assistants for marketing professionals to generate ideas, strategize better, and create more content. But to get the best results from these tools, marketers need to communicate with them clearly. It’s both an art and science to write effective AI prompts.
By Jared Frank | 7 min read
For marketers new to AI, prompting the machine can seem a bit tricky at first. However, once you understand the basic building blocks to structure your prompts well, you can guide the AI to produce high-quality content, tailored insights, and innovative ideas that match your needs. With practice, anyone can learn to write effective AI prompts.
Whether you’re looking to create engaging blog posts, detailed sales reports, or save time preparing for an annual performance review (and so many more use cases), learning to write effective AI prompts is your skeleton key. With a few well-chosen words, you can unlock a world of possibilities with AI.
Prompting is the heart of Generative AI. Prompting is like a programming language, but it’s our language – English. Where traditional programming is rigid, prompting is open-ended. Where traditional developers type out exactly what they want the computer to do, and it follows those instructions exactly every time, AI prompt engineers expect elements of randomness and inconsistencies with AI models.
So, what exactly is an AI prompt?
Key Takeaways
AI output is only as good as your prompt – the written set of instructions or question set you provide that guides an AI tool on what you need from it.
To build a good prompt, follow a simple three-step process: Prep, Prompt, Polish. In the “Prompt” step, there are six building blocks you can use to write effective AI prompts.
Never just copy/paste an AI output. The final output should always come from you, the human. Always conduct thorough quality checks before putting any content in front of clients.
AI Prompts Explained
An AI prompt is a written set of instructions or question set that guides an AI tool on what you need from it. The prompt frames the conversation with the AI, directing its focus and approach. It could be as simple as a request for a social media post or as detailed as asking for a full marketing strategy.
When you submit a prompt, the AI responds by interpreting your request based on its training and the patterns it has learned from vast amounts of data. The AI doesn’t truly “understand” your prompt in the way a person would, but it analyzes the words and context to make the best guess at what you need. The clearer and more specific your prompt is, the better the AI can align its response to your expectations.
AI output is only as good as your prompt. It takes time to practice and experiment with best practices to cultivate a good “relationship” with your AI. Ultimately, prompting should feel like a natural conversation with the AI. Users should not think of success as providing one input and letting the AI do the rest of the work. Effective prompting is a back-and-forth dialogue that ensures the AI has all the context it needs. When AI models underperform, nine times out of ten, it’s because the user has not provided adequate context.
Three-Step Process to Write Effective AI Prompts
To build a good prompt, use this three-step process: Prep, Prompt, Polish.
Prep
In this step, users create a plan for success. First, you need to define your task as simple or complex. Simple tasks are straightforward and usually require only one or two steps to execute. They can be completed with one prompt, plus some polishing. Generating ten blog post ideas, writing an email, or proofreading a slide deck are examples of simple tasks. Don’t overthink prompts for simple tasks. Simple task equals simple prompt.
Complex tasks require more advanced prompting tactics. They usually involve time-consuming, multi-step workflows and necessitate the user bring subject matter expertise to the job. AI can’t do your job for you. AI simply augments your expertise. Developing a marketing strategy or conducting a code review are examples of complex tasks.
A common misconception with AI is that work is done instantly. For complex tasks, you might spend an entire day working with AI to get the job done. But without AI, the same task might take you four days to complete. So while not instantaneous, the relative speed to completion is significant.
Rather than needing one prompt like with simple tasks, you will need multiple prompts to complete complex tasks. Don’t write a single, long, complicated prompt and expect good results. Break your task down into its subtasks, and plan prompts for each step.
Whether pursuing a simple or complex task, you should know the answers to the following questions prior to prompting so you can share this context with the AI:
- What is your objective?
- What type of output are you looking for? (an outline, table, article, etc.)
- Who is your target audience?
- What information do you need to provide the AI?
- What additional context, facts, examples can you provide?
Prompt: The Six Basic Building Blocks
Effective prompts are derived from six building blocks. Not all use cases require all six blocks, but all prompts will require at least two blocks. The six blocks are:
- Task Description – Tell the AI what you want it to do.
- Context – Provide background information.
- Role – Instruct the AI to answer from a particular point of view.
- Specific Requirements – Describe what the output should look like.
- Boundaries – Tell the AI what not to do.
- Reasoning – Ask the AI to explain its thinking to you.
Here are three examples of the six blocks in action.
Example 1:
Example 2:
Example 3:
Polish
Never just copy/paste an AI output. Marketers should always do thorough quality checks for accuracy, hallucinations, format, and perspective before putting any content in front of clients or internal stakeholders. You should also always review for cultural and emotional sensitivity. AI will never be able to use empathy as well as you.
The Human Touch Is Still Required to Write Effective AI Prompts
The final output should always come from you, the human. The powerful large language models that run AI engines have many weaknesses. They are not good at reasoning or mathematics, and they can only generate text (no images). Moreover, they are trained on old information, and they can’t search the web.
“I’ve seen a lot of hype about, oh, AI is going to replace us. We’re all going to lose our jobs,” says Edmundo Ortega, Section School instructor and partner at Machine & Partners, an AI consultancy firm. “I have seen zero indicators of any place where an AI is replacing a person fully.”
So while AI is revolutionizing work productivity across the globe, it isn’t perfect. A human still needs to initiate the work that’s done and judge the work after it’s done. And it starts with a thoughtful, well-crafted prompt. 
Are you struggling to generate high-quality AI outputs? Let’s chat.
Write to Jared at [email protected].
Disclosure: ChatGPT helped ideate the first draft of this article. The author revised subsequent drafts and contributed original copy to better reflect the intended message and voice.
RELATED CONTENT
6 Advanced Prompt Writing Tactics for Improved AI Outputs
Large language models are exciting tools, but they can lose steam with some marketers who don’t immediately experience success with them. If you want to use AI beyond basic tasks, you must move beyond basic prompts.
The First Draft: Your Business’ Starting Point for AI-Generated Content
Generative AI remains all the rage across the economy. But a lot of AI-generated content still isn’t up to par, especially early, unedited iterations. It’s not how you start, but how you finish that creates content worth publishing.