Using AI to Build WordPress Sites Is a Huge Waste of Computing — It's Dying Technology

There is a piece of advice making the rounds: point an AI coding agent at WordPress and let it “save” the platform for you — patch the plugins, wrangle the themes, keep the whole thing alive. It sounds efficient. It is the opposite. Using AI to build a WordPress site is a huge waste of computing, because you are spending some of the most capable technology ever built to prop up technology the web is already leaving behind.

The problem was never that AI can't handle WordPress. It is that WordPress is the wrong thing to pour compute into — dying technology you would not choose to build on if you were starting today.

The advice going around

Search “use AI to build a WordPress site” and you will find no shortage of people selling it as the smart move: let an agent assemble the plugins, write the glue code, and rescue a stack that has grown too complex for a human to babysit. The pitch is that AI finally makes WordPress manageable again.

But look at what that actually asks the model to do. It spends its reasoning on plugin conflicts, theme quirks, and a decade of backward-compatibility baggage — someone else's mess — instead of on your business. Making WordPress bearable is not the same as WordPress being worth building on.

WordPress is dying technology

WordPress earned its dominance in a different era of the web — server-rendered PHP, a plugin for every feature, a page builder to lay it all out. That era is receding. Modern sites are fast, purpose-built, and increasingly judged by AI search engines that reward clean structure and punish bloat. WordPress carries years of backward-compatibility weight just to keep decade-old plugins working, and that weight shows up as slower pages, heavier maintenance, and a stack that fights the very things — speed, clarity, machine-readability — that now decide whether a site is visible at all.

It will not vanish tomorrow. But it is aimed at a version of the web that is already fading, and no amount of AI changes the direction it is pointed. Automating the build just carries you deeper into a dead end.

A huge waste of computing

Here is the part the “AI to the rescue” advice skips. Every hour an agent spends reasoning about plugin soup, patch schedules, and glue code is compute burned on maintenance the platform should never have required. A very capable model spent as a systems integrator for someone else's plugin collection is close to the least valuable thing it could be doing — and you are paying for that compute, in tokens and in time, only to arrive at a fragile, high-maintenance site faster.

Contrast that with pointing the same compute at something clean. On a small, purpose-built codebase, AI spends its effort making your content sharper, your pages faster, and your site more visible to search and AI search. Same technology, radically different return. The waste isn't AI — it is aiming AI at WordPress.

Insecure at the foundation

The dying-platform problem has a security twin. WordPress is the most attacked CMS on the web, and its security is only ever as strong as its weakest plugin — a typical site runs a dozen, each written by a different author, updated on a different schedule, all trusted with the same database. Patchstack and Wordfence catalog thousands of new WordPress vulnerabilities every year, and the overwhelming majority live in third-party plugins and themes.

So anything built on top — by a human team or an AI agent — inherits every one of those weaknesses on day one. We have written before about how CMS security debt compounds. AI in the build step does not pay that debt down — it signs you up for it sooner.

Build on something worth operating

The alternative is simple: do not rescue a dying foundation — leave it. A site that is a small, clean, purpose-built codebase has a fraction of the attack surface, nothing to patch on someone else's schedule, and a structure that both search engines and AI models can read completely. That is the kind of thing worth putting an agent in charge of — because there is no plugin roulette for it to lose.

That is the whole difference. One path spends compute keeping a dying platform on life support and hopes the patches keep up. The other starts from something secure and fast, and spends its energy making your site clearer and more visible.

Where SWATS comes in

SWATS does not build on WordPress, and we do not point an agent at it and hope. We build each site as a clean, fast, purpose-built property — no plugin lottery, no theme roulette, no update-day anxiety, nothing dying bolted underneath — and then an agent operates it for you. The compute goes into your content, your speed, and your visibility in search and AI search. Not into keeping seven plugins from arguing.

And the operating part is the whole point of SWATS: you pitch us your business once, we build the site, and after that updating it is as simple as sending an email. The part you love is the only part you do.

Want to know where your site actually stands? Run the free SWATS Scorecard and see if you're visible — or invisible — to AI search.

This is the way. 🤖

— Jenaro Diaz, Founder & CEO, SWATS AI