AI-First Practices
How to use AI effectively for coding, documentation, and learning
AI is used actively across our company. This page explains how to leverage AI tools effectively while maintaining quality and accountability.
How AI Is Used
Coding
- Generate boilerplate, tests, and refactors
- Explain unfamiliar code and suggest improvements
- Debug errors and explore solutions
Documentation
- Draft README sections, comments, and docs
- Summarize long discussions or PRs
- Improve clarity and structure
Learning
- Understand new frameworks, libraries, and patterns
- Get explanations of concepts and best practices
- Explore alternatives and trade-offs
Using Page Actions
Every documentation page includes actions to work with AI:
- Copy Markdown: Copy the page content to paste into an AI chat.
- Open: Quick links to open the page context in ChatGPT, Claude, Cursor, or Scira AI.
Pro Tip
Use "Copy Markdown" and paste into your preferred AI tool with a specific question. For example: "I pasted our GitHub workflow doc. How do I create my first PR?"
Prompting Tips for Technical Tasks
- Be specific: Include language, framework, and constraints.
- Provide context: Paste relevant code, error messages, or docs.
- Iterate: Refine your prompt if the first answer is off.
- Verify: Always test and review AI output before using it.
When to Use AI vs. When to Ask Humans
| Use AI for | Ask humans for |
|---|---|
| Understanding code and docs | Decisions that affect architecture or product |
| Drafting and exploring options | Approval of changes or approaches |
| Learning concepts and patterns | Mentorship and career guidance |
| Debugging and troubleshooting | Clarification of team norms and processes |
Accountability
You are responsible for the quality of your work. AI assists; it does not replace your judgment. When unsure, ask a teammate or mentor.