Master the Perfect ChatGPT Prompt Formula (in just 8 minutes)!

How to Build ChatGPT Prompts That Actually Work

Jeff Su's guide breaks down the science behind effective ChatGPT prompting into six concrete components: Task, Context, Exemplars, Persona, Format, and Tone. Rather than offering vague tips, the video treats prompt construction as a reproducible formula that applies equally to ChatGPT and Google Bard. The underlying premise is straightforward—the better your prompt structure, the better the AI output—but the execution requires intentional design across each element.

The tutorial walks through how each component contributes to clarity and quality. A strong Task statement leads with an action verb and defines your end goal explicitly. Context uses three guiding questions to ensure you're providing relevant and sufficient background. Exemplars (specific examples) let the AI reference concrete samples rather than guess at your intent. Persona involves deciding what role or expertise you want the AI to adopt. Format focuses on how you want the final result to look. Tone addresses the emotional or stylistic quality of the response, with the video suggesting you can even ask ChatGPT itself to generate tone keywords to choose from.

Standout Strengths and Limitations of This Prompt Framework

• Modular approach: You don't need all six components every time; the framework lets you mix and match based on your specific need, making it flexible across use cases.
• Action-verb focus in Task: Starting with a verb (write, analyze, summarize, generate) immediately clarifies intent and helps avoid ambiguous requests.
• Exemplars as a quality multiplier: The video emphasizes that providing one or two real examples can dramatically improve output quality by eliminating guesswork.
• Practical tone-generation trick: Using ChatGPT itself to generate a list of tone keywords is a meta-efficient way to define voice without overthinking.
• Before-and-after comparison: The video includes a real-world example comparing good and bad prompt outputs, showing the formula in action rather than just in theory.
• Persona often overlooked: Many users skip persona definition, but the video makes clear that specifying who you want the AI to "be" (expert, beginner, writer, analyst) significantly affects results.

Who Should Use This Prompt-Writing Guide

This tutorial is most useful for anyone who uses ChatGPT or Bard regularly but feels frustrated with inconsistent or mediocre outputs. If you've ever asked an AI for help and received something generic, off-topic, or poorly formatted, the framework addresses exactly that pain point. Content creators, marketers, writers, developers, and anyone relying on AI for work will find the formula directly applicable to real tasks within hours.

It's equally valuable for people new to AI tools who want to build good prompting habits from day one, rather than learning through trial and error. The video is deliberately short (eight minutes) and example-driven, so it rewards both beginners and intermediate users without requiring prior knowledge of prompt engineering. If you're comfortable with AI tools but want to stop settling for mediocre results, this guide delivers fast, actionable structure.

Frequently Asked Questions About ChatGPT Prompt Construction

Why does the Task component need to start with an action verb?

Action verbs like "write," "analyze," "compare," or "generate" signal to the AI what kind of work you expect it to do. Without a clear verb, your request becomes ambiguous—the AI won't know if you want a summary, an analysis, a list, or something else entirely. Starting with a verb immediately narrows the scope and improves precision.

What are the three guiding questions for structuring Context?

The video references "three guiding questions" to help you decide what context matters, but the specific questions aren't detailed in the description. The principle is to provide enough background so the AI understands your situation without overwhelming it with irrelevant details. Think: What does the AI need to know about your goal, your audience, and your constraints?

How many Exemplars (examples) should I include in a prompt?

The video emphasizes that exemplars can "drastically improve" output quality, but doesn't specify an exact number. Typically, one or two concrete, well-chosen examples are sufficient to steer the AI in the right direction. More examples can help, but diminishing returns set in quickly; clarity and relevance matter more than quantity.

Can I use this formula for both ChatGPT and Google Bard?

Yes. The video explicitly presents the six-component framework as applicable to both ChatGPT and Google Bard, so the same structural approach works across these major AI platforms. The underlying logic—clearer structure equals better output—holds regardless of which AI tool you're using.

How do I decide what Tone to use in my prompt?

The video suggests asking ChatGPT itself to generate a list of tone keywords you can choose from. Instead of guessing (formal, casual, sarcastic, academic), you can prompt the AI to suggest tone options and then pick the ones that match your desired output. This meta-approach saves time and often reveals tone options you hadn't considered.