How it works
The mapper scans each sentence for capitalized entities, then searches for ~35 relationship verbs (acquired, founded, partnered with, led by, invested in...). When it finds a verb with an entity on either side, it emits a triple of subject --[predicate]--> object. No external NLP models - just regex over your prose.
Why triples beat prose for AI search
When an assistant answers a question like "who founded Anthropic?", it walks a graph of triples, not a wall of text. If your content makes relationships explicit, you make it trivial for the model to extract and cite you. Read the full entity playbook in entity SEO for the AI era.
Ship it
- Write in active voice with a named subject and object per sentence. "Apple acquired Beats in 2014" beats "the 2014 acquisition was completed".
- Wrap key relationships in schema using the Entity Schema Builder and the Organization schema generators.
- Pair with the Entity Extractor to see which entities are present, then use this tool to check that each one has at least one explicit relationship to another.