WordPress AI Lead Reveals the Truth About SEO in the AI Agent Era — It All Comes Back to Fundamentals?

Hi everyone, this is Neo.

Lately, AIO (AI Optimization) and GEO (Generative Engine Optimization) have been all anyone can talk about. A lot of independent site owners are anxious: “Now that AI agents are everywhere, is nobody going to visit my site anymore? Is my old SEO strategy completely dead?”

I totally get the anxiety. When ChatGPT, Perplexity, and even DeepSeek start handing out direct answers instead of sending traffic to your site, the traffic faucet really is being rerouted.

But James LePage, the head of WordPress’ AI team, recently shared some surprisingly sharp insights that might let you breathe a little easier — or at least make it clear where you should actually be putting your effort.

As the guy running the AI direction for the world’s biggest website-building platform (WordPress powers 43% of the internet), his take is blunt — and almost counterintuitive.

Today, Neo is going to walk you through his article and dig into what SEO should really look like in the AI agent era.

01 AI Agents Don’t “Make Things Up” — They Still Rely on Traditional Infrastructure

The first point LePage makes: AI agents still use the existing web infrastructure.

A bunch of startups are raising money hand over fist for AIO tools, as if they’ve invented some brand-new magic. But LePage points out that agents get information the exact same way as always — through existing search indexes.

“ChatGPT uses Bing, Anthropic uses Brave, Google uses Google. The web hasn’t changed — only who’s browsing it has.” — James LePage

What does that mean? It means the classic SEO signals still matter more than ever:

  • Search: agents need search to discover entities.
  • Domain Authority: agents rely on trust signals to judge whether a source is credible.
  • Links: agents hop between entities through links.
  • Content: agents need content to understand what an entity actually offers.

Neo’s take: Don’t assume the AI era means throwing out everything you know about SEO. If anything, link building and domain authority may matter more than before. AI models need to trust your data source before they’ll cite you. If your site looks like spam to a traditional search engine, it will almost certainly look untrustworthy to an AI agent too.

02 So-Called “AI SEO” Is Basically Long-Tail Keyword Optimization on Steroids

This is the sharpest point in the whole article, in my opinion.

LePage says that all the fancy tech AIO and GEO companies love to brag about, once you strip it down, is really just basic long-tail query optimization.

Why does he say that?

Because when people talk to AI, their questions tend to be extremely specific and long (like: “find me a pair of over-ear headphones for a 30-year-old programmer, with noise cancellation, under $200”). Isn’t that just what we used to call long-tail keywords?

But AI has raised the bar on content quality and structure. LePage highlights three core elements:

  1. Schema Structured Data: mark up your content in a language machines can read.
  2. Semantic Density: pack your content with substance and information — no filler.
  3. Interlinking: keep clear links between pages so AI can understand the context.

“AI intermediaries need structured, accessible content to synthesize. Clear schema, high semantic density, good interlinking. That’s the challenge most publishers are facing now.”

Neo’s take: In the old days, you could stuff keywords and fool the search engines. But AI agents are there to “extract knowledge.” If your article is mostly filler and the AI can’t pull out any solid facts after reading it, it won’t cite you. Schema markup is going from a nice-to-have to a must-have for independent sites.

03 Content for Agents: a “Briefing,” Not a “Document Dump”

LePage makes a great analogy:

Future content should read like a well-organized briefing, not a pile of documents.

Both might contain the same information, but for an agent that’s in a hurry to find answers, the difference in efficiency is night and day.

Here’s the structure he suggests:

  • Lead with the most important information (agents want direct answers).
  • Separate authoritative info from supporting info (using ranking signals).
  • Disclose details progressively (start with the summary, then point to deeper content).

The content may still be static, but as you write, always keep in mind: “This is being written for AI to read.”

Neo’s take: This circles right back to the logic of Markdown. Clear H1, H2, H3 headings, lists, bold key points. Stop writing long, rambling paragraphs with no focus. Make your content as clean and scannable as an API doc.

04 The Future Vision: From Human-Computer Interaction to Machine-to-Machine Interaction

At the end of the article, LePage paints a picture of a further future.

The current stage is: user → search/AI → synthesized information → user decision. The future stage is: user sets a goal → agent acts autonomously → agents interact → task completed.

In that future, the role of websites could change even more dramatically. A site might not even need to “look like a website” — it could just be a pile of data endpoints built for agents to call.

But before that day comes, we’re living in an in-between zone.

In this in-between zone, whoever formats their content best for agent consumption wins the race.

Neo’s Summary and Advice

After reading through James LePage’s insights, here are the takeaways Neo thinks you can act on right now:

  1. Don’t abandon traditional SEO: backlinks, authority, and technical SEO are still the foundation.
  2. Embrace structured data: check your independent site — have you added Product Schema, Article Schema, and FAQ Schema yet? That’s your “introduction letter” to AI.
  3. Raise your information density: when you write, pack in data, facts, and unique opinions. Cut the correct-sounding filler.
  4. Optimize your content format: lean into Markdown, lists, and clear subheadings.
  5. Chase long-tail traffic: think about the specific questions your users would ask an AI, and create content around those questions.

SEO in the AI era is basically turning all those standards that used to be “fine to skip if you wanted to coast” into hard requirements where “if you don’t do it, you die.”

Since the rules have changed, let’s adapt — and become the “authoritative source” that AI agents prefer.

Reference: