Creating an Enterprise-Wide View of Risks Through Deep Learning
This session will address the theme of “The Agile Auditor in the Age of Disruption” by telling the story of Employment and Social Development Canada’s (ESDC) venture into adopting machine learning technologies to transform the way risk information is analyzed by Internal Audit. The session will cover the successes and lessons learned of the venture as well as how advanced data analytics can improve risk-based audit planning, including enterprise risk management.
In the age of disruption, auditors are faced with rapidly changing environments due to advancements in technology and shifting organizational priorities, leading to significant organizational transformation. Many audit professionals are being asked to extend beyond traditional assurance roles to provide strategic insights and advice for senior management.
Auditors can better equip themselves to provide value in this age of disruption by leveraging advanced data analytics. Successful deployment of data analytics allows insights on emerging risks to be drawn quickly, and decisions by senior management to be made confidently.
At ESDC, the Internal Audit function is currently (as at December 2017) participating in an initiative that will look to explore and implement deep learning capabilities to enhance enterprise-wide risk intelligence. Deep learning is a subfield of machine learning concerned with algorithms inspired by the structure and function of the brain. One of the long-term goals of this initiative is to demonstrate the value of deep learning capabilities in enhancing risk intelligence within the Government of Canada.
More specifically, ESDC’s Internal Audit function in collaboration with the Chief Data Office have been developing a deep learning solution that will analyze risk information from across the Department and allow users to identify, visualize and analyze risk linkages within and across business lines. Furthermore, the solution will enable the recognition of trends, patterns and relationships within vast volumes of data that would be either very difficult or impossible for humans to process.
In summary, the deep learning solution will endeavour to improve the completeness, quality and horizontality of the organization’s risk-based audit planning.