The Prompt is Right!

chad

Welcome to The Prompt is Right!—the only show where AI prompting is a competitive sport and bad prompts get the boot. I’m Chad, your resident skeptic who once tried to use AI to automate my grocery list and somehow ended up with 40 pounds of potatoes.

karen

And I’m Karen, your friendly neighborhood instructional designer and AI enthusiast. We’re here to help you navigate the world of AI prompting, so you can get the best results every time. Chad, why don’t we start by talking about why good prompts are so important?

chad

Absolutely, Karen. AI is only as good as the prompts we give it. If you give it a vague prompt, you get vague results. If you give it a clear, structured prompt, you get useful output. Think of AI like an intern. If you just say, ‘Make a PowerPoint,’ you might get 30 slides of stock photos and random buzzwords. But if you say, ‘Create a 10-slide PowerPoint summarizing the impact of AI in instructional design. Use concise bullet points, relevant visuals, and a professional tone,’ now you’re getting somewhere.

karen

That’s a great analogy, Chad. And that’s where the G.O.A.L.S. framework comes in. Good AI prompts follow five key steps: Give AI a Role, Outline the Task, Add Context & Details, List the Desired Output, and Set Limits or Rules. For example, if you’re creating a cybersecurity training module, a good prompt would be, ‘Act as an instructional designer specializing in compliance training. Draft an interactive scenario-based quiz on cybersecurity risks for non-technical employees. Deliver three multiple-choice questions with rationales for correct answers. Keep each question under 20 words and avoid technical jargon.’

chad

Exactly. And when you do that, AI stops being a guessing machine and starts being a useful tool. Let’s dive into some common mistakes people make when prompting AI. The first one is vagueness. If you don’t give AI direction, you can’t expect good results. For example, if you say, ‘Write a training module,’ you might end up with a 20-slide PowerPoint of definitions with no engagement.

karen

That’s so true, Chad. Another common mistake is forgetting context. AI doesn’t know your audience unless you tell it. For instance, if you’re creating an infographic on the ADDIE model and you just say, ‘Make an infographic about ADDIE,’ AI is going to throw a flowchart at you with random colors and unreadable text. A better prompt would be, ‘Design an instructional infographic explaining the ADDIE model. Use a step-by-step flowchart with color-coded sections, simple icons, and minimal text. Ensure accessibility by using high contrast and readable fonts.’

chad

Absolutely. And the third mistake is assuming AI is perfect. It’s not. It mirrors its training data, which means biases, errors, and inconsistencies. That’s why you always have to review and refine AI-generated content. Speaking of which, let’s put our knowledge to the test with The AI Prompt Gauntlet! I’m going to throw out bad AI prompts, and you have to fix them, Karen. Ready?

karen

I’m ready! Let’s do this. First up—our instructional designer needs an AI-generated course on cybersecurity. Here’s their prompt: ‘Make a cybersecurity training course.’ Chad, that’s gonna result in 20 slides of definitions with no engagement. Fix it.

chad

Okay, how about this: ‘Create a 30-minute e-learning course on cybersecurity best practices for non-technical employees. Include interactive decision-making scenarios, real-world phishing examples, and a final knowledge check.’ Much better, right?

karen

Exactly. Now it’s clear, engaging, and actually useful. Let’s try another one. Next, our team needs a training infographic on the ADDIE model. Here’s what they asked AI: ‘Make an infographic about ADDIE.’ Chad, go ahead and fix it.

chad

Sure. ‘Design an instructional infographic explaining the ADDIE model. Use a step-by-step flowchart with color-coded sections, simple icons, and minimal text. Ensure accessibility by using high contrast and readable fonts.’ Now that’s something we can actually use in a course.

karen

Perfect! Chad, let’s talk about the role of context in AI prompts. Why is it so crucial?

chad

Context is everything. AI needs to understand who the audience is, what their needs are, and what the end goal is. For example, if you’re creating a training module for new hires, you need to specify that they are beginners. If you’re creating a module for advanced users, you need to specify that they have prior knowledge. Without context, AI is just guessing, and that can lead to content that’s either too basic or too advanced.

karen

Absolutely. Another important aspect is the impact of vagueness on AI outputs. Can you share an example of how vagueness can derail a project?

chad

Sure. Let’s say you’re creating a marketing campaign and you ask AI to ‘Write a press release.’ Without specifying the target audience, the key message, or the tone, you might end up with a generic, lifeless piece of content that doesn’t resonate with anyone. A better prompt would be, ‘Write a press release announcing our new product launch. Target tech enthusiasts and highlight the innovative features. Use an upbeat and engaging tone.’

karen

That’s a great example, Chad. Now, let’s touch on AI and ethical considerations. How do good prompts help mitigate biases and errors in AI-generated content?

chad

Good prompts help by providing clear guidelines and context. For example, if you’re creating content that involves sensitive topics like race or gender, you can specify that the AI should avoid stereotypes and use inclusive language. A prompt like, ‘Write a blog post about diversity in the workplace. Highlight the benefits of a diverse workforce and use inclusive language throughout,’ helps ensure that the content is ethical and respectful.

karen

That’s really important. Finally, let’s talk about best practices for refining AI-generated content. What are some tips for making sure the final output is top-notch?

chad

First, always review the content for accuracy and relevance. Second, test it with your target audience to get feedback. Third, refine the prompt based on the feedback. For example, if you’re creating a training module and the initial output is too complex, you can refine the prompt to ‘Simplify the content for a non-technical audience.’ This iterative process helps you get the best results.

karen

Great tips, Chad. And that brings us to our final topic—future directions in AI prompting. What do you think the future holds?

chad

I think we’ll see more sophisticated tools and frameworks for crafting prompts. AI will become even more integrated into our daily lives, and the ability to create effective prompts will become a valuable skill. We’ll also see more focus on ethical AI and transparency in how AI is used.

karen

That’s exciting. Thanks for joining us on The Prompt is Right! Share your best AI prompts with the team, and let’s keep making smart prompts for smart results. Chad, any final thoughts?

chad

Just remember—AI can be your best assistant or your worst intern. It all depends on you. See you next time!