One embodiment provides a method for generating an executable extraction program from a visual representation, the method including: utilizing at least one processor to execute computer code that performs the steps of: obtaining a validated model comprising a plurality of concepts, wherein each of the concepts comprises semantics associated with a visual representation of an object; generating at least one intermediate model object by parsing the validated model, wherein each of the intermediate model objects comprises a concept object; translating the at least one intermediate model object into executable source code; and generating an executable information extraction program from the executable source code. Other aspects are described and claimed.