We audited our own knowledge graph and found broken edges

We claim this site is built as an interconnected knowledge graph, not a pile of isolated pages. So we ran an actual audit against that claim instead of assuming it still held: grepping every page template for JSON-LD blocks and every cross-reference for whether it rendered as a real link.

It found two real defects. First, the Comparisons, Industries, and Technologies detail pages had no structured data at all — invisible to the citation-style summarization that Google AI Mode, ChatGPT, and similar systems do, despite the content being genuinely there. Second, use-case pages listed their 'typical stack' as a plain joined string of technology names with no links — a broken edge in the graph, not a missing feature, since the technology pages already existed and just weren't being pointed to.

Both got fixed the same day: TechArticle/FAQPage schema added to all three page types, and the typical-stack section rewritten to render real links to each technology's own page. We're logging it here instead of quietly shipping it, because 'we audit our own architecture' is a claim that should have a receipt.