DefinedTerm Schema
DefinedTerm schema exists for exactly one job: stating that a specific string is a term and that a specific sentence is its definition. For glossary content, nothing else in the schema vocabulary is as precisely targeted.
Its properties do real work. The name field fixes the canonical form of the term. The description carries the definition itself. Alternate names capture the variants people actually use, so a model searching for one phrasing still finds your page. Grouping entries under a DefinedTermSet tells a machine that forty pages form one coherent glossary rather than forty unrelated posts.
The strongest implementations keep three things identical: the definition in the schema, the first sentence on the page, and the meta description. When all three agree, a model has no competing versions to choose between — and the sentence it quotes is the one you wrote.
For a glossary built to be cited, DefinedTerm schema is the single highest-leverage technical addition. It is purpose-built for the exact content type AI systems reach for when someone asks what something means — and it is the difference between being paraphrased and being quoted.
A glossary entry for "backlink" carries DefinedTerm markup with the term name, a one-sentence definition, and two alternate names. Asked what a backlink is, an assistant returns that sentence with attribution. A competing page with the same information in ordinary prose is paraphrased instead — accurately, but in the model's words rather than the author's.
