Prompt Case 1
# Case Study 1 — Improving a Blog Content Prompt
## Objective
Demonstrate how weak prompts produce low-quality AI outputs and how structured prompt engineering improves results.
---
# Weak Prompt
Write a blog about productivity.
---
# Problems with This Prompt
| Issue | Explanation |
|------|-------------|
| Lack of clarity | Topic is extremely broad |
| Missing audience | Who is the blog for? |
| No structure | AI does not know format |
| No length constraint | Output length is unpredictable |
| No tone instruction | Style may vary randomly |
Prompt Health Score: **3 / 10**
---
# Improved Prompt
Write a 400-word blog post about productivity tips for remote workers.
Include 5 practical tips with short explanations.
Use a friendly and motivational tone.
---
# Improvements Made
| Improvement | Benefit |
|-------------|---------|
| Defined audience | Content becomes relevant |
| Added word limit | Output becomes consistent |
| Structured tips | Improves readability |
| Defined tone | Matches target audience |
Prompt Health Score: **9 / 10**
---
# Output Comparison
### Weak Prompt Result
Generic productivity advice without clear structure.
### Improved Prompt Result
Structured tips with actionable advice and better readability.
---
# Key Insight
Small structural changes in prompts can significantly improve the quality of AI-generated content.
This demonstrates the value of systematic prompt engineering.when to use it
Community prompt sourced from the open-source GitHub repo pranjalsharma14/promptclinic (no explicit license). A "Prompt Case 1" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
codingcommunitydeveloper
source
pranjalsharma14/promptclinic · no explicit license
more in Coding
Coding✓ tested
Senior code review (strict mode)
senior staff engineer running a merciless but fair review
Coding✓ tested
Debug by hypothesis, not by guessing
debugging partner who forms theories before touching code
Coding✓ tested
Generate tests from described behavior
test engineer who writes tests that would actually catch regressions