I’ve been spending quite a bit of time recently playing with Claude and WordPress.
The experience is genuinely remarkable. You can ask Claude to generate layouts, write content, build small apps, and connect things together in ways that would have required a developer not that long ago.
But after a few weeks of experimenting, a slightly uncomfortable thought started creeping in.
What happens when something breaks?
Because the reality is that none of these systems exist on their own anymore.
A typical workflow might look something like this:
Claude generates some content.
Another AI tool generates images.
An automation tool moves everything into WordPress.
Something else schedules social posts.
Another service analyses performance.
Individually, each tool works brilliantly.
But the moment something goes wrong, you suddenly realise how many moving parts there are.
Maybe the images don’t appear.
Maybe the formatting breaks.
Maybe the API fails halfway through publishing.
And that leads to a surprisingly tricky question.
Who does first-line support?
Do you contact:
• the AI tool
• the automation platform
• WordPress
• the plugin developer
• the hosting company
Each one will understandably say some version of:
“That part isn’t us.”
And technically they’re right.
This is the hidden complexity of what people are starting to call the agentic software world. We are stitching together workflows made up of lots of small, intelligent tools. But from the user’s perspective it all feels like one system.
When it works, it feels like magic.
When it breaks, it becomes a support maze.
Which is why I increasingly think there’s a huge advantage in platforms that own the full experience. One surface. One workflow. One place to go when something stops working.
Because the truth is, users don’t care which AI agent failed.
They just know their thing didn’t work.
And someone needs to answer the support ticket.
Leave a Reply