Machine Learning in Operational Risk

White paper - September 2019

The purpose of this white paper is to make a case for the use of advanced analytics, and specifically machine learning techniques, for operational risk management in financial firms.

It lays out the opportunities that ORX believes lie in the use of these techniques, and provides information on how they can be integrated into day-to-day operational risk management activities.

In 2017, ORX conducted a study interviewing 15 Chief Risk Officers (CROs) on the future of operational risk. Many of the interviewees identified advanced analytics (beyond capital modelling and stress testing) as a potential game changer for operational risk management. Around half earmarked it as a priority for innovation and investment over the next five years.

However, despite its huge potential machine learning has remained largely unexplored by operational risk. In a recent literature review of the application of machine learning in risk management, only six out of 50 papers focused on operational risk management.

We believe that the application of advanced analytics, including machine learning and artificial intelligence (AI), will be a core part of any future strategy for the management of operational and non-financial risk. This paper focuses specifically on the opportunities that machine learning can offer. It is intended for operational risk functions who are beginning to explore using these techniques, and want to make a business case for their application.

Based on conversations with ORX member institutions, we identified five areas in which we believe operational risk functions will benefit from the use of machine learning:

1. Freeing up valuable resources
2. Gaining deeper insights into data
3. Supporting business needs effectively
4. Gaining the skills for an enhanced level of challenge
5. Benefitting from economies of scale

Disclaimer: ORX has prepared this resource with care and attention. ORX does not accept responsibility for any errors or omissions. ORX does not warrant the accuracy of the advice, statement or recommendations in this resource. ORX shall not be liable for any loss, expense, damage or claim arising from this resource. The content of this resource does not itself constitute a contractual agreement, and ORX accepts no obligation associated with this resource except as expressly agreed in writing. ©ORX 2024

Contacts: