Enterprise-wide Risk Management (ERM) is the holistic approach to managing an organization’s upside and downside risks towards meeting its objectives. Its primary aim is to maximise risk-adjusted returns by giving consideration to the organisation’s risks and their dependencies.
Some of the traditional approaches used in risk management have been impeded by challenges such as dealing with unstructured data which limits risk management capabilities.
In a bid to advance the goal of ERM therefore, Artificial Intelligence (AI) based solutions have been increasingly deployed such as for risk identification, risk assessment and in risk management.
Using AI, current unstructured data is used to identify patterns and behaviours that then provide indications of future actions such as through advanced predictive analytics.
The increasing trend in the deployment of AI in risk management can be found in areas such as Credit Risk where the use of machine learning algorithms are used to conduct better assessments of customers’ credit histories and identify other vulnerabilities or patterns that may not have been captured. This capability through AI, aids more reliable credit scoring and the achievement of better default rates for lending institutions.
Also for market risk, the deployment of AI has aided the reduction of the risks in trading while increasing returns. Also, for fraud risk, AI models are able to analyse large data volumes, observe patterns across channels and catch potentially fraudulent activity across numerous clients all at the same time.
Major benefits accruing to organisations from deploying AI in risk management include:
• Increased focus on analytics and more proactive mitigation of losses as against the normal tendency to expend time managing the risks in operational processes
• Better identification of new and hidden risks
• Faster and more accurate risk assessments using financial and non-financial data
• New risk management approaches
• Better model risk management including back-testing and model validation
• Better risk oversight and monitoring
• Quicker and more cost-effective predictive analytics-based fraud detection across multiple channels
AI could however come with potential consequences. This could include Stakeholder Risk whereby its output could portend redundancies for certain categories of staff or with customers resistance such as in dealing with chatbots.
AI could also pose Model Risk with its ability to re-caliberate after roll-out, following the initial caliberation. It could also impact the overall risk profile of the organisation. These potential collateral or consequential risks posed by AI can be mitigated through close monitoring initiatives. They do not therefore diminish the strong transformation that AI is driving in risk management.
In general, with the deployment of AI tools in risk management, organisations achieve higher levels of risk management efficiency.
Overall benefits to the organisation include speed and time savings, lower operational costs, lower regulatory and compliance costs and then, revenue optimisation.