I had a quick back-and-forth on X this week with Matt Cromwell that I thought I would expand on a litte. He’d posted about Astro and EmDash, both of which are interesting tools. Then he said this:
I like Astro. I’m watching EmDash closely. But this is the new “Agency lock-in model”. Create a site that can only be updated via Cursor/Claude Github etc, and clients can’t touch their site themselves and keep paying agencies tons. AI enthusiasts over-estimate the common users’ capabilities or comfortability with AI tooling. Maybe one day it’ll get there, but today’s not that day.
My reply:
Every time we get a new brief in, it mentions that the marketing team want to be able to add/edit content, blocks, templates, pages etc. WordPress provides a way to do this easily. This will not change.
Matt’s point back was that it’s obvious to people who actually work with clients, but the tech-first dev audience needs reminding now and again. He’s right.
A few years ago, people were saying that the future was static sites in a git repo. Markdown files, build pipeline, fancy CDN, very fast, very modern. All of that is true. It’s also a workflow that suits people who already understand git, version control, and the command line. That isn’t most marketers. It isn’t most marketing teams. It probably isn’t most of the people who actually have to log in on a Wednesday afternoon and update a price or swap a banner before a campaign goes out.
The current AI-first variant feels pretty similar. Astro, EmDash, Cursor, Claude in your editor, GitHub pull requests for content changes. Great if you’re a developer. But if the workflow assumes the person making a change can comfortably operate a CLI, write a coherent prompt, review a diff, and merge a pull request, then you’ve designed a CMS that excludes the people the CMS is meant to serve.
This is a personal opinion, and it’s not about the technical specifics. The tools are early-stage and some will no doubt mature into something genuinely useful for non-technical users. Some won’t. And that’s fine.
But what I will say is that generally, marketing teams want to add content. They want to edit blocks. They want to build a landing page on Friday morning without filing a ticket. That requirement has been on every brief I’ve seen at Filter for years, and it isn’t because marketers are stuck in their ways. It’s because their job is to develop campaigns, not relearn the website every six months because the dev community moved on to a new favourite.
WordPress, for all its quirks, gives them that. So does Webflow. So do a handful of others. The reason WordPress still powers a meaningful chunk of the web isn’t nostalgia. It’s that someone non-technical can sit down at it and get something done.
None of this is an argument against trying new things. I run a 60-person agency and we look at new tooling constantly. We’ve got a product, PersonalizeWP, and we spend time looking at shiny products all the time. I’m always the first to upgrade as soon as something new comes along!
But there’s a gap between “this is interesting to me as a CTO” and “this is the right call for a marketing team of four trying to keep up with the campaign calendar”.
Shiny new is rarely the right answer on its own. The right answer is usually the one that respects how the people using the thing actually work.