Using process mining to gain insight of software processes presents several challenges, especially in preparing good-quality input data for the mining algorithm from the low-level data recorded by non-process-aware systems, such as version control systems. This paper proposes a process mining framework integrating AI techniques to enhance and simplify the execution of process mining tasks in this context. First, we implement a learning process to build an ontology capturing the knowledge of the analyzed processes' domain. Then this domain ontology is used in constructing an event log that enables an effective process discovery and analysis. The domain ontology plays the role of a domain expert in matching low-level events to high-abstraction and meaningful process activities. We implemented and experienced the proposed framework to extract process knowledge of e-commerce web applications development.