Banking Industry Process Mining for Conformance

Which control questions can you answer?

For account opening, loan approval, or another in-scope process, use the available evidence to investigate:

• Who approved the case? Review recorded activities and case attributes to see who performed or approved each step.
• Were the required steps completed in the right order? Compare the event sequence with the approved process to find skipped, repeated, or out-of-order activities.
• What evidence supports the decision? Review event data and case details from your source systems. The evidence available depends on the fields captured in your event log.
• Where is the control gap? Find cases where an expected control activity is missing, occurs at an unexpected point, or leads to a non-conforming path.

This analysis helps your team assess control performance. Your bank is responsible for interpreting the findings and deciding whether its requirements are met.

How does conformance checking identify deviations?

Document the expected process, then compare it with the process reconstructed from event data. The results show where observed cases differ from the standard, including missing activities, unexpected steps, and changes in sequence.

A non-conforming variant is a process path that differs from the expected flow. Grouping cases by variant shows whether a deviation is isolated or recurring. You can then review the relevant cases and available evidence instead of relying on a sample alone.

Understand conformance checking

What does a deviation and its evidence look like?

Suppose your approved account-opening process requires a review before approval. The event log shows that some cases were approved before the review was recorded.

Export case-level evidence from your analysis for review. The export can include only information present in the data you provide. It supports your audit work, but it does not certify compliance or replace your bank’s control assessment.

See what audit-ready evidence looks like

Which banking processes can you examine?

Process mining can help you examine banking processes when event data records activities and their sequence. Examples include:

• Account opening and KYC reviews
• Loan assessment and approval
• Transaction processing and payment exceptions
• Customer onboarding and service requests

For KYC process mining, look for repeated information requests, missing review steps, or unexpected handoffs, if your source data captures those events. The KYC onboarding guide for PEGA lists the activities to look for, the tables to extract, and a data template you can fill in.

Customer satisfaction in account opening and KYC

A repeated information request, a missing review step or a wait between two steps is something the applicant experiences directly, and it is also visible in the event log. Compare variants of the same process by elapsed time and by how often a case returns to an earlier step, then open the slowest cases and read what happened. That gives you the process side of a customer satisfaction question; the score itself comes from your own surveys, and the two are worth reading side by side.

The same method works for loan applications, card applications and service requests, wherever your data records elapsed time per case.

How do you investigate a control gap?

Start with the approved process and the control activities your team expects to see. Compare those expectations with actual cases, then review deviations by process variant, case, time period, or other available attributes.

For each finding, check that the event data is complete and that the case record supports your interpretation. A missing event may point to a control gap, or it may reflect incomplete logging. Confirm the cause with the relevant process and control owners before deciding what to do.

How can process mining support bank process optimization?

Bank process optimization starts with seeing where actual execution differs from the approved flow. Process mining can show where cases wait, repeat activities, or take unexpected paths. This gives operations, risk, and internal control teams a shared basis for deciding what to investigate.

ProcessMind also supports process mapping, dashboards, and KPIs. Use them to document the standard and monitor the measures your team defines. Process Simulation can help you assess proposed changes before implementation, but it does not execute them.

For a broader view of how process analysis supports improvement, read our guide to data-driven process improvement and learn how to analyze your process.

How do you get started?

Choose an account process where control evidence matters and event data is available. Define the approved sequence, required controls, and case fields your team needs to review.

Then compare the standard with actual execution. Check a small set of deviations against source records, confirm that the event log captures the relevant activities, and decide whether the findings point to a control gap, process variation, or data issue. Once the analysis works for your team, apply the same method to another process.