Please install Yoast or RankMath to use breadcrumbs.

From WordPress to LLM: The Complete AIOX Pipeline

End-to-end, what’s the actual path that a WordPress post takes to become citable by an LLM? This article traces the journey from “Publish” button to “cited in a ChatGPT answer”, through every intermediate hop.

Step 1 — You publish a post in WordPress

Standard WP workflow. Editor → Publish. The post is now live at a stable URL. WordPress’s built-in stuff fires: sitemap entry created, RSS feed updated, schema.org Article markup emitted by your SEO plugin (Yoast / RankMath / whatever).

Step 2 — AIOX Suite plugin notices

The Suite plugin hooks into save_post. On publish, it calls back to app.aioxsuite.com via a signed REST request, telling the dashboard there’s new content to process. The dashboard queues the post for processing (immediately, unless you’ve configured a delay).

Step 3 — Processing runs

The 11-stage pipeline (covered in detail in another article) reads the post, classifies it, extracts entities, infers intent, assembles a schema.org JSON-LD, merges licensing, signs, and pushes the resulting Capsule back to the Suite plugin.

Step 4 — Capsule served from the WordPress site

The Capsule is now stored on your WordPress server and served from /aiox-capsule/{slug}/. The site-wide manifest at /aiox.json is updated to list the new Capsule with its modified timestamp.

Concurrently, the Suite plugin updates the rendered HTML of the post itself: enriched schema.org JSON-LD is embedded in the <head>, TDM-REP headers are added to the HTTP response, and a discovery link points to the Capsule URL.

Step 5 — AI crawlers fetch

This is the part you don’t control directly. The major AI crawlers operate on their own schedules:

  • Perplexity: live retrieval. Reads your sitemap + Capsule manifest on demand. New content can be cited within hours of publication.
  • ChatGPT Search (OAI-SearchBot): rolling crawl + targeted re-crawl. New content typically picked up within days.
  • Gemini (via Google retrieval): same indexing pipeline as Google. Usually within days.
  • Claude: training-cycle dependent for what gets baked in; retrieval-on-demand for what’s accessible during conversation.
  • GPTBot (training): periodic crawl. Content ends up in the next training cycle, which can be weeks to months out.

Step 6 — User asks a question

Someone types “What’s the best Y for use case Z?” into ChatGPT, or Gemini, or Perplexity. The LLM either:

  • Answers from training — your AIOX-processed content was in the training corpus, so the LLM has implicit knowledge of it.
  • Answers from retrieval — the LLM runs a live search, pulls your Capsule (faster than scraping the HTML), and cites you in the response.

Step 7 — The user clicks (or doesn’t)

If the LLM cited you with a link (Perplexity always does; ChatGPT Search usually does), the user can click through to your site. That’s referral traffic. AIOX tracks it via UTM parameters.

If the LLM only used you to inform the answer without citing — also useful, but harder to measure. The AI Visibility Score is how you measure that less-tangible visibility.

How long does the full pipeline take?

  • WP publish → Capsule live: 5-30 seconds (auto-processing) or whenever you click Process (manual mode).
  • Capsule live → first Perplexity citation: 0-72 hours (depending on query freshness and topic competitiveness).
  • Capsule live → ChatGPT Search citation: 1-7 days.
  • Capsule live → baked into training: 1-6 months (next training cycle).