The primary facility in graph databases is each node explicitly contains a list of relationships it has with other nodes. The relationships are treated as first-class citizens. Neo4j has a great set of docs that introduce Graph DBs to people already familiar with RDBMSes
Another good resource is Designing Data-Intensive Applications [1]. Chapter 2 does a really good job explaining how different categories of databases relate to different data models, including examples of querying graph-like data models using `WITH RECURSIVE` compared to a query language for graph databases.
They're similar in complexity, but made easier by the fact that graph databases explicitly exist to query these types of relationships. Where someone may not know how to find their root manager node in an employee table, it would be an ideal thing to query for in a graph database. There are functions like 'shortestPath' in neo4j for example.