Open Graph for AI

Open Graph for AI recognises that the Open Graph meta tags originally designed for social media link previews have become a secondary structured data layer that AI engines also consume. When AI retrieval systems fetch a page, they parse Open Graph tags (og:title, og:description, og:type, og:image, og:url) alongside the page's visible content and Schema Markup to build a complete understanding of what the page is about.

This matters for AEO because Open Graph tags often contain the most concise, editorially curated description of a page's content. When a brand writes an og:description, they are crafting a summary specifically designed to represent the page accurately in a compact format — exactly the kind of signal AI engines value when deciding how to describe or cite a source.

The practical implications extend beyond basic metadata. For AI engines that retrieve and process web pages in real time (Perplexity, ChatGPT with browsing, AI Overviews), Open Graph tags serve as a quick-reference layer. If the og:title clearly describes the page topic and the og:description provides a concise summary, the AI retrieval system can more quickly determine relevance and extract the right framing for its response.

Open Graph for AI also includes the og:type property, which helps AI systems categorise content. An og:type of "article" signals editorial content; "product" signals commercial content; "website" signals a general landing page. These type signals complement Schema.org markup in helping AI engines understand the content's role and context.

The og:image tag, while primarily visual, can influence AI processing in engines that use multimodal models. As AI engines increasingly process both text and images, having a relevant, descriptive og:image can contribute to the AI system's understanding of the page topic.

Brands should audit their Open Graph tags with AEO in mind. Common issues include: generic og:descriptions that don't capture the page's unique value, missing og:type declarations, og:titles that differ significantly from the page's H1 (creating conflicting signals), and placeholder images that provide no meaningful context. Each Open Graph tag should be crafted as if it were the brand's elevator pitch for that specific page — because in the AI retrieval context, it often functions as exactly that.