AI is the topic of the year in the WordPress ecosystem. Every week, a new product launches with a sprinkle of AI: automated content creation (yuck!), chatbots, smart blocks, you name it. Some of these ideas are genuinely useful. Others feel like a quick “AI-washing” layer, added mostly for marketing.
But not every product needs AI.
I had the pleasure of discussing AI last week at the Web Agency Summit with Pascal Birchler from Google and Robert Jacobi of Blackwall, and realized that while all those cool features are fun, we’re not talking enough about the underlying infrastructure.
The missing layer: shared infrastructure
What we do need is to think clearly about the kind of foundation WordPress needs to support thoughtful, useful, and scalable AI features. If we want AI in WordPress to actually help people, and not just create noise, we need shared infrastructure and common standards.
One of the most promising steps in that direction is the AI Services plugin by Felix Arntz. It provides a standardized API for connecting external AI tools to WordPress, whether it’s OpenAI, Anthropic, or others. Developers can build on it without having to reinvent the wheel for every integration.
At the Cloudfest Hackathon, we explored this idea further with the MCP project. Our goal: define how AI services should work in WordPress in a way that is modular and extensible. Work like that seems to also have been happening at Automattic in this GitHub repository. That means thinking about things like user context, content history, and consent. The boring-but-essential parts that make AI useful and responsible.
Real examples: onboarding that actually helps
We don’t have to look at Lovable just for examples of what could be. We’re already seeing the impact of this kind of thinking in user-facing experiences in our own ecosystem.
Take Extendify, for example (disclosure: Emilia Capital, where I’m one of two lead partners, is an investor). Extendify helps WordPress users get started faster by using AI to combine and power design templates, intelligent recommendations, and smart onboarding. It’s not AI for the sake of AI, it’s AI that meets users where they are and helps them move forward with confidence.
Or look at the new website builder recently launched by WordPress.com (owned and run by Automattic, another Emilia Capital portfolio company), which followed closely on the heels of the Elementor AI website builder. They both use AI to ask natural-language questions and guide people to a relevant starting point. Instead of handing users a blank slate, it helps them express their goals and build from there.
Looking ahead: let’s build smarter
These examples show what’s possible when AI is thoughtfully integrated, but they also highlight the need for more shared tools, frameworks, and standards. Right now, every plugin or platform is building its own solution, from scratch. That’s not sustainable.
If we want AI in WordPress to be more than a gimmick, we have to build for it. That means investing in APIs, thinking about long-term architecture, and building infrastructure that others can rely on.
The shiny stuff is fun. But the foundational work is what will shape the future. The work that Felix did with his AI services plugin, what we did at Cloudfest on MCP for WordPress and the work done at Automattic is a very good start. Now let’s do this the open source way: together. Maybe it’s time for an AI team for WordPress?
Leave a Reply