conformance checking in banking processes using process mining: a case study

Ensuring compliance in highly regulated sectors such as banking requires organizations to continuously assess whether operational processes conform to predefined internal policies and external regulatory requirements. while process mining has been increasingly adopted for transparency and data-driven process improvement, empirical studies focusing specifically on conformance checking in real-world banking environments remain limited. this study applies a structured, five-phase methodology to assess the conformity of a loan approval process in a major iranian bank using process mining techniques. event logs containing 9,966 recorded activities across 774 loan cases were extracted, cleaned, and analyzed against a reference model derived from expert interviews and official documentation. the conformance analysis revealed that only three valid execution variants were permitted in the reference model; however, 24 variants emerged from the real execution data. although early-stage activities demonstrated full alignment, later stages involving multi-level approvals exhibited significant deviations. in total, 36.4% of cases contained at least one non-conformance, predominantly caused by delayed activity logging, missing timestamps, and manual reporting procedures rather than procedural violations. temporal analysis further showed that 31.3% of cases exceeded the mandated processing time. these findings underscore the role of process mining not only as an instrument for compliance verification but also as a diagnostic tool for identifying system-level inefficiencies. the study offers actionable recommendations to enhance workflow automation, strengthen system constraints, and improve compliance governance.