Artificial intelligence (AI) based automatic rule generation

An Artificial Intelligence (AI)-based method of automatic rule generation comprising:

• receiving user selection of a domain from a plurality of domains for generation of a domain-specific ontology;
• compiling into the domain-specific ontology, user-provided information that includes entities, attributes of the entities and relationships between the entities that are to form an hierarchical structure of the domain-specific ontology;
• providing access to the entities of the domain-specific ontology and user-selectable custom actions peculiar to the domain as elements of a drag-and-drop graphical user interface (GUI) wherein the elements include operators that comprise at least one custom action operator providing access to the user-selectable custom actions;
• enabling building of a rule flow via drag-and-drop operations on the elements, wherein the elements are dragged and dropped onto a whiteboard of the drag-and-drop GUI;
• receiving via the drag-and-drop GUI, user selections of the entities and the at least one custom action operator conveyed via the drag-and-drop operations on the elements;
• generating a visual interface upon user selection of the at least one custom action operator;
• receiving, the user selection of one of the user-selectable custom actions via the visual interface generated by the at least one custom action operator, wherein the at least one custom action operator controls at least one third-party function that transmits an input of the user-selectable custom action to at least one external system and receives processing output generated in response to the external system acting on the input;
• parsing information of the rule flow received from the drag-and-drop GUI regarding the user selections;
• determining that the rule flow complies with a set of rule requirements;
• generate code for the rule flow from the parsed information and the received processing output;
• converting the parsed information of the rule flow into at least one rule, wherein the at least one rule is a probabilistic rule and the code for the at least one rule includes code of the user-selectable custom action;
• and
• publishing the at least one rule as a web service for access by robotic process automation (RPA) procedures via a web service.

A non-transitory computer-readable storage medium comprising machine-readable instructions that implement an Artificial Intelligence (AI)-based automatic rule generation by causing a processor to:

• receive user selection of a domain from a plurality of domains for generation of a domain-specific ontology;
• receive information regarding the domain-specific ontology to be generated for the domain, the domain-specific ontology including entities, attributes of the entities and relationships between the entities;
• build the domain-specific ontology via generating code for the domain-specific ontology, the code establishing a hierarchical structure connecting the entities, the attributes of the entities and the relationships between the entities;
• enable access to the domain-specific ontology via a drag-and-drop graphical user interface (GUI), the drag-and-drop GUI including a whiteboard that receives user selections of one or more of the entities and one or more operators including at least a custom action operator via drag-and-drop operations;
• provide access to user-selectable custom actions peculiar to the domain via the custom action operator;
• receive, via the drag-and-drop user interface, visual information related to at least the user selections of the entities and the custom action operator to be used for at least one rule to be generated;
• generate a visual interface upon user selection of the custom action operator, the visual interface enables user selections of the user-selectable custom actions;
• receive, the user selection of one of the user-selectable custom actions via the visual interface-provided by the custom action operator, which controls at least one third-party function transmitting an input of the user-selectable custom action to at least one external system and receives processing output generated in response to the external system acting on the input;
• parse the visual information received via the drag-and-drop interface;
• determine that data obtained by parsing the visual information complies with one or more requirements for rule generation;
• generate code for the rule from the parsed visual information and the received processing output wherein the rule includes a probabilistic rule and the code for the at least one rule includes code of the user-selectable custom action;
• and
• publish the at least one rule to an execution server, wherein the publishing includes enabling access to the rule to one or more robotic process automation (RPA) procedures via a web service.