Prompting Is Just 7 Things You Do in Everyday Conversations

Ever stare at ChatGPT’s blank input box, wondering what magic words will make it work?

You’re not alone. The phrase “prompt engineering” makes it sound like you need specialized training. Tech forums are full of complicated frameworks with multiple steps and special syntax. Somewhere along the way, talking to AI started feeling like a technical skill you don’t have.

Here’s what changes everything: prompting is just conversation. Specifically, it’s seven things you do in everyday conversations with colleagues, friends, and family.

No new skills needed. No technical knowledge required. Just communication patterns you’ve been using your whole life, applied to a new context.

Why This Actually Matters

The biggest barrier to using AI isn’t the technology - it’s believing you need special expertise to use it well.

When prompting feels technical and mysterious, people either avoid trying or assume their results are bad because they lack some secret knowledge. Neither is true.

Once you see that prompting uses skills you already have, the intimidation disappears completely. You stop worrying about “doing it right” and start focusing on whether you’re getting useful results.

Where This Shows Up in Your Day

This matters in three places you’re probably already using AI:

• Writing emails or reports - Getting AI to draft something that sounds right and hits the right tone

• Analyzing data or information - Asking AI to spot patterns, summarize findings, or explain what’s happening

• Explaining things to others - Using AI to help you clarify concepts for your team, boss, or clients

The gap isn’t your technical skill. It’s recognizing that you already know how to do this.

The 7 Things You Do in Everyday Conversations

1) Give Clear Instructions

When you ask someone to help with something, you don’t just say “help me.” You explain what you need, why you need it, and any important details.

“Can you review this draft and check if the logic makes sense? I’m presenting it tomorrow, so focus on clarity over style.”

Same idea with AI.

Here’s what that looks like:

Vague:

Analyze this data

Clear:

Review this sales data and identify the top 3 products by revenue in Q4. Flag any products with month-over-month declines.

The second version works better for the exact same reason it would work better when asking a colleague. You’re being specific about what you want and what matters most.

Try this now: Next time you need something from AI, write your request as if you’re emailing someone capable but unfamiliar with your project.

2) Ask Good Questions

Think about the last time you asked someone for advice or information.

Good questions get useful answers. Vague questions get vague answers.

“How do I fix this?” versus “This keeps timing out when I run it on large files - what’s usually causing that?”

Here’s the difference:

Vague:

How do I clean data?

Specific:

I have customer addresses with inconsistent formatting - some include apartment numbers, some use abbreviations. What’s a reliable way to standardize these for analysis?

You’re providing enough context so the other side understands what you’re actually dealing with. It’s normal to refine your wording—everyone does.

3) Provide Context

When you brief someone on a task, you don’t just hand it over cold.

You explain the background, why it matters, and what success looks like.

“I need this summary for executives who don’t see our weekly reports. They care most about trends, not individual data points.”

Without context:

Summarize this report

With context:

Summarize this quarterly sales report for executives who don’t work with this data regularly. They care about year-over-year trends and regional differences. Keep it under 150 words.

You’ve provided: audience, priorities, and constraints.

Quick practice: Before you type anything, answer these in your head: Who’s this for? What’s the purpose? What constraints exist? What should be emphasized?

4) Iterate Based on What You Hear

Let me show you what a typical back-and-forth looks like:

You:

Help me visualize this sales data

AI: [Gives generic chart suggestion]

You:

I need to show seasonal patterns across three product lines. What chart type makes those trends clearest?

AI: [Gives specific chart recommendation with reasoning]

Multiple exchanges aren’t failure. They’re how you get to something useful. Most people need 2-3 exchanges to land on exactly what they want.

5) Show Examples

Sometimes words aren’t enough. When you want someone to understand your standard or style, you show them.

Here’s a mini-checklist for using examples well:

□ Keep a sample of formatting you like (table, report structure, tone)

□ Paste it directly into your prompt

□ Point out what specifically you want to match

Without example:

Format this nicely

With example:

Format this data like this: [paste sample]. 
Use the same column order, date format, and number rounding.

One clear example beats a paragraph of description.

6) Set Boundaries

When you delegate something, you set boundaries about what’s in scope.

• “Keep it under two pages.”

• “Don’t use technical jargon - the audience isn’t technical.”

• “Focus on Q4, not the full year.”

The key idea is specificity:

Without boundaries:

Explain regression

With boundaries:

Explain linear regression for someone who knows Excel but not statistics. Focus on what it’s used for and how to interpret results, not the math formulas. Use business examples.

Being clear about what you don’t want is often as helpful as what you do want.

7) Verify You Understood Correctly

After someone explains something, you check your understanding.

• “So what you’re saying is...”

• “Let me make sure I have this right...”

Apply the same move with AI.

After getting code:

Walk me through what this code does. 
What assumptions is it making about my data?

After getting analysis:

What are the limitations of this approach? What might it miss?

You’re accountable for the work, whether AI generated it or a person did.

---

The One-Screen Cheat Sheet

Good prompts are just clear requests with boundaries.

Copy this framework:

What I need: [Specific task] 
Context: [Who it’s for, why it matters] 
Format: [Length, style, structure] 
Boundaries: [What to include, what to skip] 
Example: [If helpful, paste a sample]

Example prompt using this:

I need a summary of this customer feedback report. 
It’s for our product team who wants to prioritize which features to build next. 
Keep it to 5 bullet points, each under 2 sentences. 
Focus on feature requests, not general praise. 
Skip any comments about pricing. 
Here’s a sample format I like: 
[paste example].

---

What Shifts When You Recognize These Skills

Once you see that prompting is just everyday communication applied to AI, several things change.

The intimidation goes away. You’re not learning something entirely new - you’re using familiar skills in a new context.

The pressure to be perfect disappears. You wouldn’t expect to give someone perfect instructions on the first try.

You focus on what actually matters: whether you’re getting useful results, not whether you’re using the “right” technique.

Your existing judgment becomes your guide. You already know when instructions are clear, when questions need more context, when examples help. That same instinct applies here.

What Gets in the Way

“I’ve seen those complicated prompting frameworks with special formatting”

Those exist for specific advanced situations. But they’re not required for effective prompting. Clear communication gets you most of the way there.

“My prompts don’t always work even when they seem clear”

This happens in regular conversations too. What seems clear to you isn’t always clear to someone without your context. When this happens, add more background or try explaining it differently. It’s normal, not a sign you’re doing it wrong.

“Shouldn’t I learn all the technical features and special syntax?”

You don’t need to know all of Excel’s functions to use it effectively. Learn what you need as you go, based on what you’re actually trying to do.

Your 3-Prompt Starter Pack

Copy and adapt these for your actual work:

For writing:

Draft an email to [recipient] about [topic]. 
Tone should be [professional/casual/direct]. 
Key points to cover: [list 2-3 points]. 
Keep it under [length].

For analysis:

Review this [data/report/document] and identify the top 3 [patterns/issues/opportunities]. 
For each one, explain what it is and why it matters. 
Audience is [who they are].

For explaining:

Explain [concept] to someone who understands [related thing they know] but hasn’t worked with [the new thing]. 
Focus on practical application, not theory. 
Use an example from [their context].

Start Here

Pick one thing you need to do with AI this week. Could be drafting an email, analyzing some information, or explaining something.

Use the one-screen framework above. Fill in each part before you hit send.

Compare the result to what you’d normally get. Notice what’s different.

The goal isn’t perfection. The goal is recognizing that clear communication is something you already know how to do.

Til next time,
Donabel