By Antoine Riesser, founder, Animam.ai · Updated May 18, 2026 · 16 min read

AISEO — Agentic / AI SEO — is the discipline of making your WordPress site discoverable, citable, and operable by AI agents (ChatGPT, Claude, Perplexity, Gemini, plus the wave of vertical agents). It’s not just structured data. It’s not just llms.txt. It’s a coherent set of 12 signals that, taken together, let AI agents find you, parse you, cite you with attribution, and increasingly interact with your site on behalf of their users.

Why AISEO matters for WordPress in 2026

Three forces are pushing every WordPress publisher toward AISEO right now:

  • AI Overviews and Answer Engines are eating search. Google AI Overviews, Bing Copilot, Perplexity, ChatGPT Search and Claude.ai have shifted the way users consume answers. A growing portion of queries never reaches a classic SERP — the answer is synthesized from web sources, with citations. If your WordPress site isn’t structured to be cited, you’re invisible to that traffic.
  • Agents now act, not just answer. An agent doesn’t just summarize your page — it can book an appointment, place an order, fill a form. WordPress sites without an agent layer (MCP server, agent-friendly endpoints) are read-only to this new traffic. Animam’s plugin makes WordPress write-friendly to agents.
  • The Model Context Protocol (MCP), published by Anthropic in November 2024, became the standard. Claude Desktop, claude.ai, Cursor and a growing list of agents auto-discover MCP servers. A WordPress site that exposes an MCP endpoint enters the agent ecosystem directly, not just through indexed pages.

The 12 signals of agent-ready WordPress

An audit-ready model. Each signal is independently testable. Together, they cover what AI agents look for when discovering, parsing and acting on a WordPress site in 2026.

1. /llms.txt at the site root

A plain-text manifest describing your site for LLMs, inspired by robots.txt but oriented toward semantic discovery. Tells AI agents what your site is about, what’s relevant, where the API lives, and what to cite. Animam’s plugin generates and serves /llms.txt automatically based on your published posts and pages.

2. /.well-known/agent.json

The canonical discovery file for AI agents. Declares the site’s agent capabilities (channels available, MCP endpoint URL, supported tools), the operating entity, contact, and privacy posture. Animam emits this automatically with the right entity metadata pulled from your WordPress settings.

3. /.well-known/llmfeed.json

A structured feed of your latest content optimized for LLM consumption — title, summary, canonical URL, publication date, language, topic taxonomy. AI agents can subscribe to this feed for incremental updates instead of re-crawling. Animam refreshes the feed on every post publish/update via the save_post hook.

4. /.well-known/webmcp.json

The MCP discovery file announcing your Model Context Protocol endpoint. Tells compatible agents (Claude Desktop, claude.ai, Cursor) where to connect and what scope to request via OAuth 2.1. This is the signal that turns a WordPress site from a read-only resource into an agent-operable surface.

5. AI fetcher firewall configuration

Cloudflare and other WAF layers now block AI fetchers (Claude-User, ChatGPT-User, Perplexity-User, etc.) by default. Your WordPress site needs a custom rule to allow these UAs. Without it, AI agents that try to fetch your page get a 403 and silently drop you from citation candidates. Animam ships a recommended Cloudflare rule for the 10 known AI UAs.

6. JSON-LD Organization schema on every page

Tells agents who runs the site: legal name, URL, logo, contact, social handles. Critical for E-E-A-T signals. RankMath or Yoast handle this on WordPress; Animam’s plugin double-checks it and flags missing fields via the admin chatbot.

7. JSON-LD Article schema with named author

Every blog post should ship Article schema with an author of type Person (not Organization). LLMs cite humans more readily than corporate bylines, and E-E-A-T scoring favors verifiable individual expertise. Animam’s admin chatbot flags posts missing a Person author and suggests templates.

8. JSON-LD FAQPage on Q&A blocks

The single most effective schema for AI citation. FAQPage with proper Q&A pairs gets pulled verbatim into AI Overviews and Perplexity answers. Animam’s mu-plugin pattern (post meta _animam_faq_jsonld + injector) lets you ship FAQPage on any page type, not just structured FAQ blocks.

9. JSON-LD BreadcrumbList for navigation

Helps agents understand your site’s information architecture. Improves citation context (the agent knows your post is in the “alternatives” section vs “tutorials”). Animam emits BreadcrumbList automatically based on the WP page hierarchy.

10. MCP server endpoint with OAuth 2.1 + PKCE

The endpoint that turns your WordPress site into an agent-operable surface. With Animam’s plugin, every tenant gets an MCP server at /wp-json/animam/v1/mcp with OAuth 2.1 + PKCE. Compatible with Claude Desktop, claude.ai, Cursor and any MCP-aware client. 17 tools available: corpus search, post creation, SEO audit, scheduling, payment, etc.

11. Agent detection (X-Agent-Model header)

When an AI agent (or another LLM) hits your chat endpoint, you should detect it and respond differently — structured JSON envelope instead of streaming SSE, machine-readable tool metadata, no human-only welcome screens. Animam detects the X-Agent-Model header and switches mode automatically. This is the A2A (Agent-to-Agent) layer.

12. Hreflang + canonical hygiene

For multilingual sites, hreflang annotations and proper canonical URLs prevent agents from citing the wrong-locale version. Polylang and WPML handle this; Animam’s admin chatbot has a “hreflang audit” tool that flags missing or broken tags.

How Animam injects all 12 signals by design

The promise of the Animam WordPress plugin: install once, get the full 12-signal stack without manually wiring each file. Here’s what gets activated on plugin activation:

  • Files generated automatically: /llms.txt, /.well-known/agent.json, /.well-known/llmfeed.json, /.well-known/webmcp.json. Refreshed on every save_post.
  • JSON-LD coverage: Organization (from site identity), Article with Person author (when post has a Polylang author), FAQPage (via post meta + mu-plugin), BreadcrumbList (from page hierarchy).
  • MCP server live: /wp-json/animam/v1/mcp endpoint exposed, OAuth 2.1 + PKCE configured, 17 tools available. Bearer secret encrypted AES-256-GCM in WP options.
  • Agent-to-Agent envelope: X-Agent-Model header detected; structured JSON response instead of SSE when an LLM hits /chat endpoint.
  • Discoverability Coach in wp-admin: a dedicated section that audits your site’s 12 signals and flags missing ones with one-click fixes.

The admin chatbot as your AISEO coach

Animam’s most underrated feature for AISEO is the admin chatbot inside wp-admin. It’s not the front-end agent visitors talk to — it’s a private assistant for the site editor. 16 admin tools cover the AISEO workflow:

  • wp_seo_audit — Scans every published post: missing focus keyword, missing meta description, missing meta title, missing featured image, images without alt text, meta description length outside the 70-160 range.
  • wp_set_seo_meta — Auto-detects RankMath or Yoast and writes meta title / description / focus keyword in one call.
  • wp_set_image_alt_bulk — Fix up to 30 missing image alt attributes in one call. Critical for AI image understanding.
  • wp_create_post / wp_update_post / wp_publish_post / wp_schedule_post — Programmatically create AISEO-optimized post drafts following your house template (Person author, structured headings, FAQ schema).
  • wp_search_posts — Find posts matching a theme/topic for refresh decisions (“which of my posts is closest to the new AI Overviews query trends?”).
  • wp_rest_call — Escape-hatch for any REST API operation when no dedicated tool exists yet.

Workflow in practice: open wp-admin, type “audit my agent-readiness” in the Animam Assistant. The chatbot runs wp_seo_audit + checks the discoverability files + verifies the MCP endpoint health + suggests fixes. Apply the suggestions with one click. The agent-readiness score goes from 4/12 to 12/12 in an afternoon.

The AISEO content patterns that actually get cited

Beyond the 12 signals, what matters for actual LLM citation in 2026:

  • Short, direct, answer-first openings. The first sentence of your post should answer the implicit query. LLMs extract the lead paragraph more reliably than buried answers.
  • Named author with verifiable expertise. “Edited by the Animam team” gets less citation than “By Antoine Riesser, founder Animam.ai” with a clickable About page linking to a real human.
  • Explicit dates. Date visible in the byline (not just in metadata). “Updated May 2026” is a citation booster for queries about a date-sensitive topic.
  • Structured FAQ at the bottom. 8-20 questions, each with a short direct answer in the first sentence + context paragraphs after. This is the AEO weapon.
  • Internal linking with descriptive anchors. “Read our top 10 WordPress AI plugins ranking” works better than “click here” for AI context-building.
  • Honest disclosure of competitive position. “This comparison is edited by Animam, which ranks first” with credit to competitors — beats pure self-promotion 5:1 on Perplexity citation rates.

Measuring AISEO performance

How do you know AISEO is working? Three signals to track:

  • Referrals from AI assistants in your analytics. Look for traffic with referrer chat.openai.com, claude.ai, perplexity.ai, gemini.google.com, copilot.microsoft.com. These visitors arrived because an AI cited or recommended your page.
  • Direct queries to your MCP server. Animam logs every MCP call to the tenant. When Claude Desktop or Cursor connects, you see it in the audit log. Growing volume = agent ecosystem is using your site.
  • Citation appearances. Search your URL on Perplexity, ChatGPT Search, Claude.ai. If your domain shows up in citations for topical queries, you’re winning AISEO.

Frequently asked questions

What’s the difference between SEO, AEO and AISEO?

SEO targets classic search engines (Google, Bing). AEO (Answer Engine Optimization) targets answer engines that summarize web sources (Google AI Overviews, Bing Copilot, Perplexity). AISEO is broader — it covers AEO plus the agentic dimension: how AI assistants act on your site, not just cite it. AISEO includes MCP, agent-to-agent envelopes, and tool exposure, which traditional SEO/AEO don’t address.

Is AISEO replacing SEO?

Not yet — classic search still drives the majority of WordPress traffic in 2026. But AI assistants are growing 30-50% year-over-year and AI Overviews appear on an increasing share of Google queries. The sites preparing for AISEO now will dominate when the shift accelerates. The good news: AISEO signals overlap heavily with SEO best practices (structured data, semantic HTML, real authors), so you don’t have to choose.

Do I need to write content differently for AISEO?

A little. Front-load answers in the first paragraph. Use named authors with bios. Add FAQ sections with direct one-sentence answers. Quote your sources explicitly. Most of these are good content hygiene anyway, but they tip the citation rate noticeably.

What’s an MCP server and why does my WordPress need one?

An MCP server (Model Context Protocol, published by Anthropic in November 2024) lets external AI agents discover what your site can do and call its tools. With Animam’s plugin, your site exposes an MCP endpoint at /wp-json/animam/v1/mcp. Power users of your site can connect Claude Desktop or Cursor to it and ask the agent to take actions on your site (book a meeting, search the corpus, generate a quote). Without MCP, your site is read-only to the agent ecosystem.

Will allowing AI fetchers cannibalize my traffic?

Some traffic moves from clicks to in-AI summaries, yes. But the alternative — blocking AI fetchers — gives you zero AI Overview visibility and zero Perplexity citations. The net effect on revenue depends on your model: for content monetized by ad impressions, the trade-off is harder; for content monetized by lead gen, contact form fill, or product purchase, AI citations drive more qualified intent than vanilla search clicks.

How do I know which AI fetchers to allow?

The current “good citizens” list at May 2026: Claude-User (Anthropic), ChatGPT-User (OpenAI), GPTBot, Perplexity-User, PerplexityBot, Google-Extended (Google AI training), Gemini-Deep-Research, Bingbot (with Copilot suffix), Applebot-Extended, FacebookExternalHit (when Meta AI fetches). Animam ships a Cloudflare rule that whitelists these 10 UAs.

Does the Animam admin chatbot really fix all 12 signals automatically?

The 4 files (llms.txt, agent.json, llmfeed.json, webmcp.json) and the 4 JSON-LD types (Organization, Article, FAQPage, BreadcrumbList) — yes, on plugin activation. The MCP server endpoint — yes, generated on first config. The Cloudflare rule — no, you need to apply it manually (we provide the JSON to copy-paste). The hreflang hygiene — flagged by the audit tool, fix is one-click. The agent detection layer — auto-active on all tenant endpoints.

What’s the difference between Animam and just adding RankMath + a Schema Pro plugin?

RankMath covers signals 6-9 (classic JSON-LD) very well. Schema Pro covers similar ground. Neither covers signals 1-5 (the .well-known files + AI fetcher firewall) or signals 10-12 (MCP, agent detection, hreflang audit). Combining RankMath + Animam is the recommended stack: RankMath handles the classic SEO + structured data, Animam handles the agentic layer (MCP, .well-known files, A2A) and the admin chatbot.

Can I test my agent-readiness score before installing Animam?

Yes. The free service isitagentready.com scans any URL and returns a level 1-5 score against the 12-signal model. Try it on your current WordPress site to see where you stand. After installing Animam Free and running the Discoverability Coach, re-scan to confirm the lift.

Is AISEO a marketing buzzword or real?

Both, honestly. The acronym is recent and the field is still consolidating its vocabulary (AEO, GEO, AI SEO, agentic SEO all roughly overlap). But the underlying signals — MCP, JSON-LD, .well-known files, structured FAQ — are documented standards adopted by major AI companies. Whatever you call the field, the 12-signal stack moves citation rates measurably. The buzzword will settle; the signals won’t.

How long does it take to make a WordPress site fully agent-ready?

With Animam: under an hour for the technical signals (plugin install, configure, validate). For the content patterns (named author bios, FAQ sections, answer-first openings), expect a few hours per high-priority page. For a 50-page site, plan a 2-3 day refresh sprint. The compounding payoff in citation rates kicks in after 4-8 weeks of crawl + reindex cycles.

What about non-WordPress sites — does AISEO apply?

Yes, the 12 signals are platform-agnostic. Animam covers WordPress with a plugin, but the model itself works for Next.js, Shopify, Webflow, or static sites. The harder part on non-WordPress stacks is the admin chatbot layer — it’s much easier to drop a plugin into wp-admin than to graft an in-platform coach into other CMS dashboards.

What’s the cost of getting AISEO wrong?

Invisibility to a growing share of traffic. By mid-2026, ~30-40% of high-intent queries in many verticals are answered without a click — directly inside AI Overviews, ChatGPT Search, Perplexity. If your competitor’s WordPress site is cited and yours isn’t, you lose the lead. The cost of inaction compounds monthly.

Where to go from here?

Three concrete next steps: (1) Run an isitagentready.com scan on your current site to get a baseline. (2) Install Animam Free (50 conversations/month, free) and let the Discoverability Coach generate the 4 .well-known files. (3) Read our companion ranking top 10 WordPress AI plugins 2026 to see how Animam compares to the other plugins in this space.

Sources and further reading

Try Animam — install the agent-ready plugin in 5 minutes

Free plan: 50 conversations/mo, 1 agent, official WordPress plugin with the full 12-signal stack pre-configured. Install the plugin →

Comments, corrections, additional signals to consider? Email [email protected] — this is a living document, updated as the AISEO field evolves.

Related: MCP for WordPress in 2026: the protocol is solved. The agent UX isn't.

The full AISEO cluster