How to Engineer Prompts like a Pro | How to AI #2

Step-by-Step Guide

Step 1: Access Gemini AI

1. Go to Google's Gemini Platform

Step 2: Use This Prompt

Copy and paste this optimized prompt:

Generate a prompt to update the comprehension of text, factually correct the information in a given text and improve the structure of the text while while maintaining the tone and soul of the text. This text is supposed to be posted on LinkedIn and we want to structure it for maximum interaction and virality.

Step 3: Add Context Before You Run It

Before pasting your draft, add three quick context lines so the model understands your intent:

1. Who you are writing for (e.g. founders, operators, marketers).
2. The tone you want (e.g. authoritative, practical, conversational).
3. The action you want readers to take (comment, click, DM, subscribe).

The more specific your context is, the more usable the output becomes.

Step 4: Ask for Structured Output

Request output in a reusable format so editing is faster. For example:

• Hook options (3 variants)
• Main body (short + long version)
• Call-to-action options (2 variants)
• Suggested hashtags (5 relevant tags)

This gives you alternatives quickly instead of a single one-shot response.

Step 5: Validate Before Publishing

Do a quick quality check before posting:

• Remove generic filler words.
• Confirm claims are factual and specific.
• Keep sentences short and skimmable.
• Ensure the CTA matches your business goal.

If needed, re-prompt with: "Shorten this by 20%, keep meaning intact, and make it clearer for first-time readers."

Practical Prompt Template

Use this format for repeatable results:

You are a senior content strategist. Rewrite the draft below for LinkedIn.
Audience: [target audience].
Goal: [engagement / leads / thought leadership].
Tone: [tone].
Constraints: Keep it under [X] words, preserve original meaning, improve clarity, and remove fluff.
Output format: 3 hooks, 1 refined post, 2 CTA options, 5 hashtag suggestions.

Common Mistakes to Avoid

• Asking for "better content" without audience context.
• Mixing multiple goals in one prompt.
• Ignoring structure (no hook, no CTA, no format).
• Publishing first output without refinement.

Why This Works

Prompt engineering works best when you treat prompts as instructions, not keywords.
Clear role, audience, constraints, and output format make results consistent.
With this approach, you can build a repeatable writing workflow that saves time while keeping your brand voice intact.

Need Help?

If you face any issues with this micro AI hack, email us at hello@thisweekinai.club

---

This guide is part of our AI Hacks series - quick, actionable solutions for everyday problems.