Coding

Coding use cases have AI generate, debug, port, refactor, and explain code. This primitive extends far beyond traditional software engineering — it includes anyone who needs to create scripts, formulas, queries, or macros as part of their work. A marketer writing an Excel formula, an analyst creating a SQL query, an engineer writing a microservice — all coding use cases.

Coding is also the primitive where AI agents are most mature: modern coding agents can plan an approach, write code, run tests, debug failures, and iterate with minimal human guidance.

Coding is one of six use case primitives identified in OpenAI’s Identifying and Scaling AI Use Cases, adapted here to be platform-agnostic.

Agentic Coding methodology

For a structured methodology to go from idea to shipped feature using AI coding agents, see Agentic Coding — a 7-step lifecycle with Vision Briefs, PRDs, test-driven development, and automated code review.

When to apply this primitive

Use Coding when:

• The deliverable is something that executes — code, scripts, formulas, queries, or macros
• You need to translate a business requirement into a technical implementation
• You’re debugging, refactoring, or porting existing code
• Non-technical team members need technical artifacts (Excel formulas, SQL queries, simple scripts)

Not the right primitive when:

• The main output is a written document — that’s Content Creation
• You’re analyzing data and the insight is the deliverable, not the analysis code — that’s Data Analysis
• You’re creating a pipeline that runs autonomously — that’s Automation, though coding may be a step within it

Worked use cases

• Agentic Coding

  Codingskillagentprompt

  When I'm building a new feature, I want a structured AI-assisted lifecycle so I go from idea to shipped code without losing context.

Related

• AI Use Cases Overview — the full library, searchable and filterable by primitive
• Agentic Coding — the full 7-step feature development lifecycle
• Product & Engineering — PRDs, user stories, acceptance criteria, and the software development lifecycle
• Agents — autonomous coding agents that plan, write, test, and iterate
• Context — providing codebases, documentation, and examples