Characteristics of a Relation

A relation is filled in like a node. Depending on the profile (Essential, Basic or Expert), some information is available or not.

Main information of the relation

Characteristic	Type	Req.	Profile	Description	Role for the AI agent
SOURCE	List	Yes	All	Source node. E.g. EMPLOYEE	Identification of business relations
TARGET	List	Yes	All	Target node. E.g. SKILL	Identification of business relations
NAME	String	Yes	All	Relation identifier. E.g. holds_skill	Identification of business relations
SHORT DESCRIPTION	String	Yes	All	Short business definition. E.g. Employee holds a skill.	Identification of business relations
CONFIGURATION	String	No	All	Bidirectional or not (valid in both directions) — all profiles. Transitive or not (if A → B and B → C, then A → C) — Basic and Expert profiles	Navigation and reasoning on the graph
CONFIDENTIALITY	List	No	Basic and Expert	C1 public access (default), C2 intermediate access, C3 secret access, C4 top secret access	Hide confidential relations according to the chosen level

Semantic information of the relation

Characteristic	Type	Req.	Profile	Description	Role for the AI agent
LONG DESCRIPTION	String	No	Basic and Expert	E.g. holds_skill links an Employee to a Skill they currently possess. This relationship captures the fact that an employee has demonstrated or acquired a given competency, which can be used to assess their fit against Job Profile requirements, identify training needs, or find employees for a specific project or task. An Employee can hold multiple Skills, and a Skill can be held by multiple Employees.	Fine understanding of relations
SYNONYMS	String	No	Basic and Expert	E.g. holds_skill → has skill, possesses skill, is skilled in, competency of, masters	Fine understanding of relations
BUSINESS RULES	String	No	Basic and Expert	E.g. An Employee can hold the same Skill only once (no duplicate holds_skill relationships between the same Employee and Skill). A holds_skill relationship must always carry a proficiency_level attribute — it cannot exist without one. If an Employee leaves the company, all their holds_skill relationships must be archived rather than deleted, to preserve historical skill data.	Fine understanding of relations

Main information of a relation attribute

Characteristic	Type	Req.	Profile	Description	Role for the AI agent
NAME	String	Yes	All	Attribute name. E.g. proficiency_level of the holds_skill relation	Identification of relation attributes
TYPE	String	Yes	All	int, str, float, bool or date	Identification of relation attributes
SHORT DESCRIPTION	String	Yes	All	E.g. proficiency_level is the level of mastery of a skill held by an employee.	Identification of relation attributes

Semantic information of a relation attribute

Characteristic	Type	Req.	Profile	Description	Role for the AI agent
LONG DESCRIPTION	String	No	Basic and Expert	E.g. proficiency_level indicates an employee’s level of mastery of a skill, on a standardized scale. Used to assess the fit between an employee’s skills and a Job Profile’s requirements, as well as to guide training needs.	Fine understanding of relation attributes
SYNONYMS	String	No	Basic and Expert	E.g. proficiency_level → skill level, competency level, expertise level, mastery level	Fine understanding of relation attributes
BUSINESS RULES	String	No	Basic and Expert	E.g. The value of proficiency_level must be an integer between 1 (beginner) and 5 (expert). A level of 5 can only be assigned after validation by a manager or an external certification. The level must be reassessed at least every 2 years.	Fine understanding of relation attributes

Relation languages

Characteristic	Type	Req.	Profile	Description	Role for the AI agent
DATA LANGUAGES	List	No	Basic and Expert	Languages in which the relation’s data is stored	Interpretation of the data language
TRANSLATIONS	List	No	Basic and Expert	Translation languages to display the relation name	None, display only