Intelligent internal linking (which no standard CMS does)
What was done
We designed an automated internal linking system that exceeds the limits of traditional CMS by relying on AI and semantic embeddings. To ensure a consistent link structure, we implemented automatic inbound and outbound flow counters, coupled with a pruning algorithm limiting density to 6 links per page to preserve link juice. This architecture allows for the instant identification and correction of orphan pages by suggesting relevant connections based on terroir or grape variety via vector analysis of content.
On the technical side, we automated the detection and cleaning of dead links while isolating transactional areas from editorial calculations to avoid diluting semantic relevance. The administration interface now includes integrated SEO diagnostics and parameterizable publication templates via optimized Markdown/HTML editors. Security has also been reinforced with granular database role management, preventing any privilege escalation during content management by the team, while ensuring perfect relational integrity between product sheets and blog articles.
Client benefit
The client now has a true structured editorial graph where each article reinforces the estate's global theme. This automation eliminates the chaos of manual links and natively optimizes technical SEO by removing navigation dead ends. The editorial team gains total autonomy, transforming complex management into a fluid workflow where contextual relevance naturally guides visitors and crawlers toward the most consistent products.