WordPress is not dead. It still powers a very large part of the public web, and its own AI project is adding the building blocks needed for assistants to work with WordPress sites.
We left anyway.
That is a less exciting headline than declaring an entire platform finished. It is also a more useful story.
Mark and Helen used WordPress across Collab365 for years. The main site, sales pages, checkout flows, member areas and several publishing projects all accumulated their own themes, plugins and exceptions.
WordPress helped us build the business. The problem was that the system we had assembled no longer fitted the way we wanted to work.
AI exposed our editing tax
Our normal editing loop had become ridiculous.
We would find a block inside a page builder, copy its text into an AI tool, explain the audience and offer again, paste the result back, repair the formatting and then discover that another version of the same claim lived elsewhere.
The AI was capable. The surrounding system did not know enough about our business to use it safely.
That distinction matters. A product can have AI features without your content becoming structured, connected or safe for an agent to update.
The installed base is both strength and constraint
The W3Techs WordPress survey still puts WordPress at roughly two-fifths of measured websites. The exact percentage moves, which is why the sensible conclusion is not tied to one decimal point: WordPress remains dominant.
WordPress is also moving. Its official MCP Adapter work uses the Abilities API so approved AI clients can discover and call defined WordPress capabilities. That directly contradicts the idea that WordPress is simply ignoring AI.
But a modern core does not automatically modernise every old site.
Our site had years of page-builder markup, plugin settings, shortcodes, redirects, account logic and content duplication. Much of the difficult migration work came from our implementation, not WordPress itself.
That is the honest lesson. A flexible platform lets you build almost anything. It can also let you accumulate almost anything.
Why we still moved
We wanted one content system that could answer four questions reliably:
- What does this page claim?
- Which source proves it?
- What other content depends on it?
- What can an AI assistant change without breaking something important?
Building that around our existing stack felt harder than creating a smaller system shaped around those questions.
That choice does not generalise to every business. If your WordPress site is stable, properly maintained and easy for your team to edit, a migration may create cost without creating value.
If it depends on an ageing tangle that nobody can confidently explain, the case is different.
Do not migrate because somebody says WordPress is dead. Map your content, integrations, permissions, redirects, forms, analytics, checkout and recovery plan first. Then compare the cost of leaving with the continuing cost of staying.
We did that comparison and left. WordPress did not fail the test for everyone. It failed the test for the business we were becoming.