Process Mining at National Transport Insurance (NTI)

 

Background

About. This case study is the result of a collaboration between Queensland University of Technology (QUT) and National Transport Insurance (NTI), Australia’s market leading provider of insurance for heavy commercial vehicles. The study applied process mining techniques to a dataset provided by NTI’s Underwriting department, which had implemented an Operational Model change in October, 2014.

Objective. This case study aimed to reveal improvements to process management and performance subsequent to the Operational Model change through an analysis of role-based handover patterns.

Key Questions. Of interest to NTI were questions such as (i) have tasks been appropriately allocated to the roles?, (ii) is the distribution of tasks, by the number of tasks (i.e. workload) equitable across roles?, and (iii) have human resources been correctly assigned to roles?

Data. The dataset was exacted from NTI’s claims management systems and comprised a set of transaction records and anonymised employee role appointment information. The event log included all transactions (quotes and policies) recorded in the period 1st July 2013 to 31st July 2015. There were 45 roles and 120 employees in the event log.

Approach. In this case study, various existing process mining tools and techniques were used including DISCO (commercial) and ProM (open source). The analysis included task distribution by roles, hand-over analysis between roles, and performance comparison of different roles. The techniques (ProM plug-ins) that were used in this study include the “Dotted Chart Analysis”, “Mining Resource Profile”, “Social Network Analysis”, and “Inductive Visual Miner”.

Results and Impact

  • There were significant reductions (up to 50%) on various case durations observed after the date of Operational Model change.
  • Following the Operational Model change, the distribution of tasks across roles became more balanced.
  • Task allocation became more specialized/targetEd to roles according to skills required to complete the task.
  • One objective of the Operational Model change was to reduce the amount of intra-role handover behaviours. This study revealed that, after the Operational Model change, there were still instances of intra-role handover behaviours, but for the nine different roles analysed for the Motor Fleet policy type, only two roles exhibited significant intra-role handover behaviours.
  • Inter-entity handover behaviours occurred in nearly 70% of cases of External delivery mode.
  • When compared to cases not involving Inter-entity handovers, cases involving Inter-entity handovers took longer to move from quote to policy and had a larger number of events per case but had a higher conversion rate (21% as compared to 15% for cases not involving Inter-entity handovers).

Publication

An article entitled as “Discovering Role-based Handover Patterns Using Event Logs: An Australian Case Study” has been prepared based on this research.