elliott richmond

In the rapidly evolving landscape of web development, AI tools are promising to revolutionise how we code. But can these tools truly deliver value for WordPress developers? This post explores a hands-on experiment with Cursor AI, testing its capabilities in creating a custom WordPress block, a practical, real-world development task.

The Challenge: Creating a Custom Icon Block

WordPress’s block editor offers many built-in blocks, but sometimes you need something custom. In this case, we needed a reusable icon block that would:

• Allow selection from a suite of SVG icons
• Support alignment options
• Include colour customisation
• Offer dimension controls

A seemingly simple but practical component that would enhance a standard WordPress 2025 theme layout.

First Attempt: AI Without Context

The first test involved asking Cursor AI to “create a custom WordPress block that lets the user insert icons” without providing any additional context or guidelines.

The results were underwhelming:

• The code was placed in the wrong folder structure
• No package.json was generated
• The code wasn’t set up as a proper WordPress plugin
• The build process was missing

While parts of the code were functional (the block.json structure was reasonable), the overall implementation was flawed. Without understanding how custom blocks work, a developer might struggle to fix these issues.

Key Takeaway: AI can generate code, but without clear instructions, the results may be basic and require significant reworking.

Second Attempt: AI with Enhanced Context

For the second attempt, three critical changes were made:

1. Added Cursor Rules – A file containing WordPress-specific coding principles and best practices
2. Used Create Block Package – Created a proper plugin structure using npx @wordpress/create-block
3. Provided More Detailed Instructions – Specified the need for SVG icons, alignment, colour settings, and dimension controls

The result was dramatically different:

• The AI modified the existing structure correctly
• Added appropriate attributes in block.json
• Implemented proper controls for icon selection, alignment, colours, and dimensions
• Created an extendable system for adding more icons

The code generated was clean, maintainable, and followed WordPress coding standards. The block worked as expected, allowing icon selection, colour changes, size adjustments, and alignment options.

Extending the Implementation

To test flexibility, a simple request to “add a location icon to the suite of icons” was made. Cursor AI immediately identified the appropriate file (icons.js) and added the new SVG icon without disrupting the existing structure. After a quick build, the new icon was available in the block editor.

Key Lessons for WordPress Developers

1. Context is crucial – Providing AI with your project structure, coding standards, and clear requirements dramatically improves output quality
2. Use existing scaffolding – Starting with a proper structure (like @wordpress/create-block) helps AI understand your project better
3. Review and validate – Always understand what the AI-generated code does before implementing it
4. Think of AI as a team member – Collaboration produces better results than simply accepting whatever code is generated

Beyond Block Development

While this example focused on creating a custom block, the same principles apply to other WordPress development tasks:

• Refactoring existing plugins and themes
• Writing custom queries and template parts
• Creating custom post types and taxonomies
• Developing admin interfaces

Getting Started with Cursor AI for WordPress Development

To replicate this approach:

1. Install Cursor AI
2. Set up Cursor Rules for WordPress (available at cursor.directory)
3. Use standard WordPress scaffolding when possible
4. Provide clear, detailed instructions
5. Review and refine the generated code

Conclusion

AI tools like Cursor can significantly boost productivity for WordPress developers when used properly. The key lies in providing context, using existing structures, and critically reviewing the output. AI doesn’t replace the need for development knowledge, instead, it augments your capabilities by handling repetitive tasks and suggesting implementations.

When given the right guidance, AI can help you create cleaner, more maintainable WordPress code in less time. As these tools continue to evolve, they’re set to become an essential part of every WordPress developer’s toolkit.

What are your thoughts on using AI for WordPress development? Would you trust AI to write your WordPress plugins? Share your experiences in the comments!