Here’s a small example.
I needed to check if I could access a project in Atlassian Jira via their API. Someone had shared the API URL with me, but I didn’t remember the exact curl syntax or headers.
So I dicatated this prompt for an LLM:
I need to run a curl command against the Atlassian Jira Cloud API to check if a project exists and retrieve its details. Specifically, I want to query:
https://domain.atlassian.net/rest/api/3/project/slug
Requirements:
curl -s -w "\nHTTP Status: %{http_code}\n" \
-u "YOUR_EMAIL@domain.com:YOUR_API_TOKEN" \
-H "Accept: application/json" \
"https://domain.atlassian.net/rest/api/3/project/SLUG" \
| jq .
Explanation:
Why did this work so well? Because the prompt was clear, specific, and asked for a concrete deliverable. Not just “help me with Jira API” but step-by-step requirements + expected output.
That’s really the trick: sometimes writing a good AI prompt is just writing a good spec.
A Note on Ethics and AI Use: Transparency is important. For this article, I used AI tools to augment my discussion and explore phrasing, as well as to assess SEO performance and readability. While AI helped refine ideas and highlight optimization opportunities, all insights, examples, and analysis are the product of my own experience and judgment. AI served as a support tool, not a replacement for critical thinking or human perspective.