A knowledge graph is a representation of data as simple entities, or nodes, interconnected by semantic relations. Within PeopleRisk, every person in your company will be represented as one of those entities, with an arbitrary number of intrinsic properties, like its name, contract type, or more personal HR aspects, etc. Every asset, every access right, the fact that this person has a position in an organizational chart or works in a specific office or country, is represented as other entities linked to the person by relations.

Knowledge Graphs are intrinsically closer to the way our brain organizes knowledge – as interconnected conceptual elements that can be used in multiple different contexts. It is thus much more agile and evolutive than traditional relational databases, designed for stable processes with few changes.

The role of our underlying CogTL engine is to maintain this knowledge graph, adapt it in realtime whenever a data changes, and re-evaluate all rules or calculations affected by this change. As an example, let us say that the country of domiciliation of an employee changes: CogTL will then automatically re-evaluate all compliance rules that make use of this country of domiciliation, typically cross-border rules or regulations.

But there’s more: as we now have a model of knowledge that federates multiple elements of knowledge coming from multiple sources – sometimes maintained in silos – we can now perform multiple statistical analytics on the model, and discover correlations, outliers, or potential anomalies in an autonomous way. For example: a person that has access to an HR resource but does not belong to HR, or a new temporary employee with much more accesses than the others, etc.