Prompting Basics

Recall: AI Predicts, Doesn’t Understand

In Lesson 2, we learned that AI generates responses by predicting likely next words. This means what you write directly shapes what you get back.

A vague prompt leads to a generic response. A specific prompt leads to a tailored response.

Today, you’ll learn exactly how to write prompts that get results.

By the end of this lesson, you’ll be able to:

Structure prompts using the RTCF framework
Write prompts that are specific enough to get useful responses
Apply the “good prompt” checklist to any request

The RTCF Framework

Every effective prompt contains four elements. We call this RTCF:

Component	What It Does	Example
Role	Tells AI what expert persona to adopt	“You are a senior data analyst…”
Task	Specifies exactly what you want done	“…analyze this sales data and identify trends…”
Context	Provides background information	“…for a retail company with declining Q4 sales…”
Format	Defines how you want the output	“…present findings as a bullet-point summary with a recommendations section.”

You don’t always need all four, but including them makes your prompts dramatically more effective.

RTCF in Action

Let’s transform a weak prompt into a strong one:

Before (weak):

“Help me write a cover letter”

After (with RTCF):

“[Role] You are a career coach who has helped hundreds of professionals land jobs at top tech companies.
[Task] Write a cover letter for a Senior Product Manager position at Spotify.
[Context] I have 6 years of PM experience at smaller startups, led a team of 5, and launched 3 products with $10M+ revenue. The job posting emphasizes data-driven decision making and cross-functional leadership.
[Format] Keep it under 300 words. Use a confident but not arrogant tone. Structure it as: hook, relevant experience, why Spotify, call to action.”

The second prompt will generate a cover letter that’s actually useful.

The Specificity Principle

Here’s the core rule: The more specific your prompt, the more useful your response.

Vague prompts force AI to guess what you want. It usually guesses wrong—or plays it safe with generic content.

Examples of Adding Specificity

Vague	Specific
“Write an email”	“Write a follow-up email to a client who missed our meeting yesterday”
“Explain machine learning”	“Explain machine learning to a marketing manager who needs to understand it for a vendor pitch”
“Give me feedback”	“Review this paragraph for clarity and suggest 2-3 specific improvements”
“Make it better”	“Make this more concise—cut it from 500 to 200 words without losing key points”

Quick Check

Which of these provides more specificity?

A: “Write a blog post about productivity”

B: “Write a 600-word blog post about the Pomodoro technique for remote workers who struggle with focus. Include 3 practical tips and a personal anecdote.”

(Answer: B—it specifies length, topic, audience, structure, and tone)

The Good Prompt Checklist

Before sending any important prompt, run through this checklist:

✅ Is the task clear? Could someone else understand exactly what you’re asking for?

✅ Is the audience/context defined? Does AI know who this is for and why?

✅ Is the format specified? Do you know what the output should look like?

✅ Are constraints included? Length limits, tone requirements, things to avoid?

✅ Is there enough context? Does AI have the background info it needs?

If you’re missing any of these, your prompt probably needs work.

Common Prompt Patterns

Here are three patterns you’ll use constantly:

Pattern 1: The Expert Request

You are a [specific expert role].

[Task: what you want done]

Context:
- [relevant background]
- [constraints or requirements]

Format: [how you want the output]

Pattern 2: The Transform Request

Take the following [input type]:

[paste your content]

Transform it into [output type] that:
- [requirement 1]
- [requirement 2]
- [requirement 3]

Pattern 3: The Analysis Request

Analyze the following [content type]:

[paste content]

Specifically:
1. [What to look for]
2. [What to evaluate]
3. [What to recommend]

Present your analysis as [format].

Try It Now: Practice Exercise

Take this weak prompt and rewrite it using RTCF:

“Help me with my resume”

Consider:

What role should the AI take?
What specific task do you need?
What context is relevant (your experience, target job)?
What format should the output be?

Write your improved version before continuing.

Example improved version:

“You are a tech recruiter who has reviewed thousands of resumes for FAANG companies.
Review my resume (pasted below) for a Senior Software Engineer role at Google. I have 5 years of experience, primarily in Python and distributed systems.
Provide:
Three strengths that will appeal to Google recruiters
Three areas to improve with specific suggestions
Any red flags or missing elements
Be direct and specific—generic advice isn’t helpful.
[Resume would be pasted here]”

Key Takeaways

Use RTCF (Role, Task, Context, Format) to structure effective prompts
Specificity is the #1 factor in prompt quality
Run the Good Prompt Checklist before important requests
Practice transforms knowledge into skill

Up Next

In Lesson 4, we’ll explore how to use context and examples to guide AI behavior even more precisely. This is where prompting starts to feel like a superpower.