Structured data fields are by definition an abstraction of legal language. When in doubt, use citations, the AI chat, and the actual contract text to validate the structured data against your own interpretation. Overview
The Entities extraction identifies the contracting parties in the document and creates entity records in your CRM. Only entities directly involved in the transfer or assignment of music rights are extracted — entities that are merely mentioned or referenced in passing (e.g., a collecting society named in a clause) are not.
Entity Roles
Each extracted entity is assigned one of three roles in the agreement:
Assignees
The entity receiving rights under the agreement. This is typically a label, publisher, or distributor. An agreement usually has one assignee.
Assignors
The entities granting rights under the agreement. These are typically artists, writers, producers, or rights holders. An agreement can have multiple assignors.
Administrators
A third-party business entity explicitly appointed to administer publishing income or rights (e.g. a sub-publisher or collection agent named with an appointment or “collects on behalf of” clause). This role is reserved for a party beyond the assignor and assignee. Administrators are extracted and stored, but the contract-detail entity cards show only assignees and assignors — administrators surface in the CRM and other party views.
For each entity, the extraction captures:
What’s Shown on the Card
The entity card on the contract detail page shows: name, address, and contract relations. A More expandable section reveals any linked artists, writers, and codes & identifiers associated with that entity on this contract. Representatives and other extracted data are stored on the entity’s CRM profile and can be viewed by clicking the entity name.
Automatic Entity Merging
When entities are extracted, Royaltyport automatically checks whether the entity already exists in your project. This prevents duplicates when the same company or individual appears across multiple contracts under slightly different names.
How Merging Works
The system uses a three-step approach:
- Exact match — first checks for an entity with the exact same name. If found, the extraction links to the existing entity immediately.
- Similarity search — if no exact match, generates a normalized representation of the entity (stripping legal suffixes like Ltd/Inc/BV, lowercasing, normalizing contact details) and searches for similar entities using vector embeddings.
- AI evaluation — if similar candidates are found, an AI model evaluates whether they’re truly the same entity. It matches on legal suffix variations, trading names, capitalization differences, and minor typos. It does not match different geographic entities, subsidiaries, or different legal persons.
Parent Entity
When entities are merged, the entity with the most contract references becomes the parent. All other variations point to it. This ensures your CRM has a single canonical record per entity, regardless of name variations across contracts.
The entity shown on a contract’s detail page is always the entity as extracted from that specific contract. The merging happens in the background to keep your CRM clean. Click the entity name to see the canonical CRM profile with data consolidated from all contracts.
Editing Entities
Click the edit icon to:
- Correct entity names or details
- Add missing identifiers
- Link entities to existing CRM contacts
- Update addresses or contact information
