How to Deploy the Semantic Infrastructure on your Website

Publishing JSON-LD for your main entities is only part of the job. Once the knowledge graph is live, the information in your content, your structured data and the graph itself has to agree. This lesson explains why alignment matters so much and how inconsistency quietly undermines the work you have already done. If AI systems see conflicting details, they lose confidence in who or what you are.

What this lesson covers

• The principle that content and entity data must always align
• Four dimensions of consistency: names, properties, bidirectional links and a single source of truth
• How to keep one canonical name for each entity across every page
• Why dates, addresses and URIs need identical formats everywhere
• A worked example comparing inconsistent and consistent entity data
• How consistency affects LLM confidence and citation behaviour

The lesson treats consistency as something you check across four dimensions. Names come first: one canonical name for each person or organisation, used on every page, with no mix of titles, initials and post-nominals. Properties come next: the same date, address and URI formats wherever they appear. Relationships must be bidirectional, so if a person works for an organisation, the organisation should list that person in return. Finally, all of this should be governed by a single source of truth, such as a configuration file or database based on the entity blueprint built earlier in the course. That record tells everyone the correct spelling and format for each attribute.

A short example makes the risk concrete. When a home page, an about page and a blog post each describe the same expert with a different name format and a different date format, language models may read them as three different people. The result is lower confidence, fragmented authority and inconsistent citations. When every page references one entity record with a canonical name and an ISO-formatted date, the same systems see one unified expert. The lesson also notes a practical benefit: a single source of truth gives you more control over what you publish in both your content and your structured data.

This lesson sits in the Knowledge Graph Integration and Optimization module, between choosing an implementation approach and extending the graph to other platforms and maintaining it. It is useful for anyone responsible for publishing or governing brand information, including SEOs, content editors and developers, who need a simple standard for keeping entity data clean after launch.