SEO used to be one job. In 2026, it’s four, and almost no WordPress plugin is doing all of them well.

Open any WordPress site and there’s a fair chance you’ll find an SEO plugin doing… something. A meta description here, a sitemap there, a colored bullet telling you a post is “OK.” For years, that was enough. Optimize a title, hit “Add focus keyword,” watch the indicator turn green, ship.

That world is over. The way search actually works has fractured, and most plugins are still solving 2018’s problems.

Update — 2026-05-18: Google published its official AI Optimization Guide on 2026-05-15 and ended most of the AI-SEO speculation industry in one document. This article has been rewritten to align with it: chunkability and llms.txt sections removed (Google explicitly says they don’t help), Pillar 4 reframed around AI Overview cannibalization detection — the genuinely new measurable signal Google’s guide implies. Full breakdown: What Google’s May 2026 AI Guide Actually Says (And What We Changed in SEO Engine).

New to AI search? The interesting part starts at Pillar 4. For the developer angle on how AI agents now talk to WordPress directly, read How To: Connect Claude Code to WordPress with MCP.

What Changed

Five years ago, “SEO” meant getting your meta tags right and making Googlebot happy. There was one search engine that mattered, one set of signals to optimize for, and the plugin market reflected that.

Today, three things have shifted:

• Google’s AI Overviews eat clicks invisibly. A page that ranks #3 can still bleed traffic when Google summarizes the answer above your link. Ahrefs and Pew measured 47% to 60% CTR loss on affected queries. Google does not tell you which queries are hit.
• Search Console has the data your plugin is missing. Most SEO plugins still link you out to Google’s dashboard. The actual answers (which posts to fix first, which queries are close to ranking, which queries are being AI-Overview’d) live one OAuth click away from your WordPress admin.
• AI bots crawl the web. GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, each has its own pattern, and most plugins can’t even tell you they came by. This doesn’t affect Google rankings, but it’s a real signal about who’s reading your content.

The Four Pillars of WordPress SEO in 2026

Here’s the model. Each pillar is a separate job; doing one well doesn’t help you with the others.

1. Perfect technical SEO. The boring infrastructure that breaks invisibly.
2. Cleaner, better content. Written for human readers — which, per Google’s May 2026 guide, is also what AI Overviews reward.
3. Visibility on Google. Knowing what’s actually happening in search results.
4. AI awareness. Tracking which AI bots crawl your content, and detecting when Google’s AI Overviews are eating your clicks.

Most SEO plugins do (1) and a piece of (2). The market has not caught up to (3) and (4), and the gap is widening every month. Let’s go through each.

Pillar 1: Perfect Technical SEO

Table stakes, and the part where bloat is most embarrassing.

You need clean meta tags, valid Open Graph cards, an XML sitemap that updates correctly, a robots.txt you can edit without SFTP-ing into your server, structured data that doesn’t break, and an og:type that says website on your homepage and article on your posts. (A real bug we shipped a fix for recently. The kind of thing technical-SEO plugins ship by accident.)

SEO Engine handles all of this without fourteen settings screens. The sitemap is auto-generated with WooCommerce, Polylang, and WPML awareness. Schema.org markup is emitted where it should be — for rich results, not because it helps with AI search (Google’s guide is explicit: schema is not required for AI). OG tags are correct in every context. One robots.txt editor with one box and a save button.

The argument here isn’t “we have these features.” Every SEO plugin claims they do. The argument is what’s not there: no upsell modal when you save robots.txt. No “Pro feature” badge on the sitemap. No fifty-step setup wizard. The plugin lets you ship and gets out of the way 🙂

That’s what technical SEO should actually mean: invisible when it’s right, immediately fixable when it’s wrong.

Pillar 2: Cleaner, Better Content

For a decade, content SEO meant “did you put your focus keyword in the H1?” That metric is dead. Google’s May 2026 AI Optimization Guide is unusually direct about what works now: a unique point of view, first-hand experience, write for humans, integrate quality images and video, be indexed, be crawlable. Google’s own contrast example is striking: “7 Tips for First-Time Homebuyers” (commodity, recycled, won’t get cited) vs. “Why We Waived the Inspection & Saved Money” (first-hand, citable). Same topic, completely different fate.

SEO Engine’s Content Clarity score measures three things — the ones that actually map to human readability (and therefore, per Google’s guide, to AI search visibility too):

• Structure. Are H2 and H3 headings present and well organized for skim-reading?
• Lists. Is at least one key fact in a bullet or numbered list?
• Clarity. Are sentences a readable length, with quotable phrasing?

Each signal contributes to a single 0 to 100 score. Plus a separate originality and first-hand POV check, scored strictly on the commodity-vs-first-hand axis Google’s guide names as the single most important content factor. Generic “7 tips” lists land in the 40s. First-hand experience pieces land in the 70s and above.

The suggestions are concrete, not abstract. Instead of “Readability: 67,” you get “This reads like a generic tip list. What’s your actual experience with the topic? A single specific story would lift this above the citation threshold.” The writer, or an AI agent acting on their behalf, can act on it directly.

This is where the Magic Fix loop matters. SEO Engine’s Magic Fix can take a post, identify the failing signal, propose an improved version, and apply it with one click. The loop is: AI finds the problem, AI proposes the fix, human approves, ship. No eight-tab dashboard. No thirty-minute audit ritual.

What we deliberately don’t do: score “chunkability,” suggest writing in a special “AI-friendly style,” or recommend chunking content for AI extraction. Google’s May 2026 guide explicitly states none of that helps. SEO Engine used to score chunkability; we removed it the week the guide came out. Honest beats marketing.

Pillar 3: Make Sure Pages Are On Google

Here’s where most SEO plugins quietly hand you off to Google’s own dashboard and call it done. Some link you out; some don’t even do that.

SEO Engine’s Search Console integration is different. After OAuth (one click, shares credentials with the Analytics setup most users already have), you get a dashboard tab that answers four questions:

• How is the site doing? Clicks, impressions, average position, with period-over-period deltas, color-coded.
• What should I fix first? “Quick Wins” splits into buckets: posts ranking #5 to #15 with strong impressions (close to the first page), high-impression posts with weak CTR (title or meta fixes), hidden gems (high CTR, low visibility), and the new AI Overview Suspects bucket (top-5 rankings with cratered CTR — the signature of Google summarizing your answer above your link). Each opportunity names the specific post, the specific search query, and the estimated click lift.
• What’s working? Top pages, top queries, with WordPress post matching for Polylang and WPML multi-domain sites.
• What’s happening with this specific post? A per-post “pulse” returning top queries, click trends, and position changes in one call.

The entire module is also exposed via MCP, which means an AI agent (Claude, ChatGPT with MCP, or your own) can ask SEO Engine “what are this week’s quick wins?” and get a structured, actionable answer in milliseconds. That’s the layer most plugins haven’t even imagined yet.

This pillar is the difference between “my SEO plugin says I’m doing fine” and “my SEO plugin knows what’s working and tells me what to fix.”

Pillar 4: AI Awareness

Here’s the part where the AI-SEO industry has been selling you snake oil, and where Google’s May 2026 guide drew a clean line.

What does NOT help (per Google, explicitly):

• llms.txt files (Google does not fetch them; adoption is ~6%)
• AI-specific markup or “AI-friendly” content style
• Chunking content for AI extraction
• Writing for every query variation (this is scaled-content spam policy territory)
• Schema markup as an AI signal (it’s for rich results, not AI Overviews)

What does help:

• The same SEO that’s always worked — first-hand content, crawlable pages, indexed properly, written for humans.
• Making sure your main content isn’t blocked behind JavaScript that AI extractors can’t process. SEO Engine flags this automatically.

What is genuinely new — and where Pillar 4 actually lives — is measurement. Google’s AI Overviews can eat 47% to 60% of your clicks on affected queries (Pew, Ahrefs). Google does not tell you which queries are hit. SEO Engine ships:

• AI Overview Suspects detection (Pro). Identifies queries where you rank top-5 but CTR cratered below 35% of expected. That’s the AI Overview signature. The category does not promise a fix — there is no fix that beats AI Overviews. The honest move is to know which queries are hit and diversify.
• AI-bot tracking. Visibility into which AI crawlers (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Googlebot variants) hit your site, which pages they read. This doesn’t affect Google rankings, but it tells you who’s reading your content beyond Google.
• MCP server. Your SEO data is reachable by AI agents directly. They can audit your site without ever opening a browser tab.

That’s the honest Pillar 4. Not “optimize for AI” — because there is no separate AI search to optimize for, per Google. Measure the damage AI Overviews are doing to your clicks, and know when AI bots are reading you.

Why Lightweight Beats Feature Breadth

The bloated SEO plugin is a recognizable shape. You install it; you get a dashboard with fourteen menu items, three upsell modals, a setup wizard that takes thirty minutes, a “premium feature” badge on basic functionality, and a daily nag email. Every version adds more.

There’s a logic to that approach. Feature breadth equals perceived value equals paid-tier conversions. But the cost is invisible until it isn’t: the plugin slows your site, confuses your editors, and makes you afraid to update.

The alternative isn’t fewer features. It’s focused features that compound. SEO Engine doesn’t try to be every SEO tool ever invented. It does four things (the four that actually matter in 2026) and connects them to each other and to AI. The Content Clarity score feeds into the Magic Fix loop. Search Console quick wins suggest specific edits. AI Overview Suspects lives inside Quick Wins as its own bucket. Each module is small; the system is intelligent.

That’s what “lightweight plus AI-aware” actually means. Not minimalism for its own sake. Compound utility from a small, careful surface.

Questions People Ask

What changed in WordPress SEO between 2018 and 2026?

In 2018, “SEO” meant making one crawler (Googlebot) happy with keyword density, meta tags, and link structure. In 2026, ranking on Google is still the core job — but now you also need to measure when Google’s AI Overviews are eating your clicks, track which AI bots crawl your content, and write content that holds up to Google’s published originality bar. The fundamentals haven’t changed; the measurement surface has.

What are the four pillars of modern WordPress SEO?

Technical SEO (clean meta tags, valid OG cards, working sitemap, schema markup for rich results), content built for human readers (first-hand experience, structure, lists, clarity — the things Google’s May 2026 AI guide names as what actually works), Google visibility (Search Console integration that tells you what to fix), and AI awareness (AI-bot tracking plus AI Overview cannibalization detection).

Is keyword density still important in 2026?

No. Google moved past keyword density as a primary signal years ago, and Google’s May 2026 AI guide explicitly warns against targeting every query variation — that falls under their scaled-content spam policy. What matters now is first-hand point of view, clear structure, and writing for humans.

Should I write content in a special “AI-friendly” style or chunk it for AI?

No. Google’s May 2026 AI Optimization Guide is explicit: “There’s no requirement to break your content into tiny pieces for AI to better understand it.” Google’s systems understand nuance, synonyms, and meaning across multi-topic pages. SEO Engine used to score “chunkability”; we removed it the week the guide was published. Write for human readers; that’s also what AI Overviews reward.

Does llms.txt help my site rank in AI search?

No. Google’s May 2026 guide states explicitly: “You don’t need to create new machine readable files, AI text files, markup, or Markdown to appear in generative AI search.” Adoption of llms.txt is around 6%, and OpenAI and Anthropic don’t fetch it either. If you have one, you can delete it — it changes nothing. SEO Engine used to ship an llms.txt editor; we removed it.

How is AI search different from Google search?

Per Google’s May 2026 AI Optimization Guide, AI Overviews and AI Mode are not a separate search engine — they use the same Google index as regular Search. So “optimizing for AI search” is not a distinct activity. What is different is the user experience: AI Overviews show a summary above the blue links, which can absorb 47% to 60% of clicks on affected queries (Pew, Ahrefs). That’s a measurable cannibalization risk, but the SEO that helps you appear in either is the same SEO.

Do AI bots like ChatGPT and Claude read the same things Google does?

They read the same HTML. ChatGPT, Claude, and Perplexity each have their own crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) and citation logic, separate from Google. SEO Engine tracks which of those bots visit your site — it’s a useful signal about who’s reading you beyond Google, but it does not affect Google rankings.

What is MCP and what does it have to do with SEO?

MCP (Model Context Protocol) is the open standard that lets AI agents talk directly to your tools. SEO Engine exposes its analytics and audit data via MCP, so an AI agent (Claude Code, ChatGPT with MCP, etc.) can ask “what are this week’s SEO quick wins?” and get a structured answer in milliseconds, without you opening a browser.

Does Yoast or RankMath support AI search?

Both ship AI-assistive features for writing content. Neither ships AI-bot tracking, AI Overview cannibalization detection, an MCP server, or originality scoring aligned with Google’s May 2026 guide. They cover classic SEO well and are adding 2026 polish at the edges; SEO Engine was built around the new measurement surface from the start.

Are AI Overviews really eating my clicks?

If you rank in the top five for any query Google decides to show an AI Overview for, very probably yes. Pew measured ~47% CTR loss; Ahrefs measured ~60%. Google does not flag affected queries in Search Console. SEO Engine’s AI Overview Suspects feature finds them by signature — queries where you rank top-5 but CTR sits below 35% of expected. There is no fix that “beats” AI Overviews; the play is to know which queries are hit and diversify.

Jordy’s Take

The thing I keep telling people is this: don’t pick your SEO plugin by feature checklist. Pick it by the questions it can actually answer. The good ones tell you what to fix next, in plain English, on a real page you wrote. The rest hand you a dashboard and wish you luck.

SEO Engine is young. It’s not the most famous SEO plugin, and it won’t be for a while. But it was built for the way search works now, and when Google published its official AI Optimization Guide in May 2026, we rewrote the parts of the plugin that didn’t match. That’s the bet we’re making: honest beats marketing 🙂

In Short

• Modern WordPress SEO has four pillars: technical, content, Google visibility, AI awareness.
• Most plugins do one and a half of them.
• If your SEO plugin can’t tell you which AI bots crawled which posts this week, it’s measuring 2018.
• If it can’t tell you which of your top-5 queries are being eaten by AI Overviews, it’s measuring 2018.
• llms.txt, chunkability scoring, and “AI-friendly markup” don’t help — Google’s May 2026 guide says so explicitly.
• SEO Engine is free on WordPress.org. Search Console integration and Magic Fix are in Pro; everything else (sitemap, schema, OG, content scoring, AI-bot tracking, MCP) is free.

Curious how SEO Engine stacks up against Yoast and Rank Math? Read our honest 2026 comparison.

Try SEO Engine on one post. See the Content Clarity score, click through the suggestions, and form your own opinion. Free, no signup, ten minutes 🙂