“What can go wrong will go wrong”. We have often heard this sentence, in fact first stated by “Murphy’s Law” that everyone in the operations field knows very well. In today’s technological advancements, the question one should ask is, Can we pre-mortem the operational processes?

“What can go wrong will go wrong - unless you identify potential problems earlier”

There are many risks or business hazards that have the potential to damage the business processes, reputation, assets, revenue, profits, etc. The risk of damage or loss is due to some failed internal processes or inadequate controls. Such hazards are difficult to plan for and hence result in risks. Identifying such risks well in advance and preparing a mitigation plan for them so that the damage is either prevented or restricted to a minimum forms the backbone of any proactive operational risk management framework.

Under the Operational Risk Framework (ORF), risk is identified through data collected on incidents, internal audits, external audits, regulatory notices, etc. Risks are then analysed and prioritised so that a firm action plan can be prepared and implemented to mitigate those risks. Thus, it generates an Operational Risk Issue (ORI) that has an accountable owner (usually the one who generates it - typically senior executives of the company) and a risk administrator (who ensures the end resolution work). The end goal is to implement an actionable and sustainable control to minimise the risk of damage in case of operational hazards. The controls are termed as Key Procedural Control if on the ORF impact scale the risk has higher priority (high probability of damage in a short period) or as Operational Risk if the control is not mitigating key risks. The process of risk identification, especially in manufacturing operations or heavy machinery/asset-dependent business, is usually reactive based on incidents, audits, or historical experiences. Sometimes, it is also performed proactively depending on the experiences of individual experts or tacit knowledge.

There is a growing demand to reduce risk response time for various risk mitigation processes, and thats driving increasing focus on IIoT-enabled machinery assets and attracting deployment of analytical tools for risk analysis and management. AI-enabled analysis solutions can identify potential failures in advance so that suitable countermeasures can be deployed to avoid losses. Such analytical tools hugely support decision-making by surfacing hidden insights from a pool of data across different sources. AI can thus help ORI owners to make sure that a sustainable solution is implemented in a quick response.

Imagine a global manufacturing enterprise that possesses multiple manufacturing plants across continents and has deployed data gathering techniques to collect real-time data on assets. Multiple sources can provide either historical or real-time data. For example, on a sensor-driven platform data is collected in real-time, most of the companies store data on files/ledgers in a database to be used later, or publicly available datasets used for environmental analysis, and so on. Being a data-rich company is an advantage but the enterprise can rip the benefits only if it squeezes out real, actionable insights from that data and use that information to fulfil business needs/goals. Only then, the real proactive risk management strategy can be implemented. Using AI-enabled analysis, such operational data on machines, equipment, and plants can be processed in real-time to find patterns or insights, and to predict failures ahead of time. This valuable information can then be used in ORF to alter the business processes, implement sufficient controls, and prepare workaround solutions. The forecasting of raw materials, breakdown spares would be more accurate and based on evidence of need before time. This would bring down the cost of inventory holding and provide greater operational efficiency. The biggest beneficiaries are senior management at the business executive level who can visualise the situation across geography, across plants, and make well-informed decisions.

Another example is the insurance industry that works on the fundamental principle of risk assessment. According to market research, the global insurance analytics market will be USD 20 billion by 2026 with a 15% CAGR during this period. The insurance sector has seen tremendous demand but at the same time has put unusual pressure on the service provider. Along with the higher sales, there are pressures due to higher claims, market uncertainty, and low-interest rates. Insurance providers must leverage advanced analytics to not only make the customer experience hassle-free but manage associated risks effectively.

The use of analysis software application is for the decision support and it can perform analysis online as well as offline. It could continuously monitor the state of any asset using collected operational data and either match it against a known normal state to find deviations. However, innovative applications are using new approaches such as pattern recognition across multiple signals to determine the non-normal state of the asset. For complex processes and data, the usual ‘signature detection’ method becomes inadequate, and ‘asset behaviour analysis’ may work better. The system or asset behaviour can be analysed while it is interacting with its surroundings in performing its task. That means not only machine data is important but its surrounding environmental parameters are also important to determine its behaviour. The challenge is analysing all associated signals coming from a variety of sources, in different formats - together. This kind of data analysis is possible through finding patterns of signals collectively and determining individual signal contributions to provide explainability feature. Click here to explore such an application.

For asset insurance companies, such predictive insights are nothing but a valuable tool to pre-empt the cost of risks in the underwriting procedure. It would enable insurers to predict fair pricing, manage their capital efficiently, detect frauds, and keep a check on the inherent uncertainties in the manufacturing industry. The artificial intelligence-based analysis enables insurers to leverage the power of predictive insights to manage customer life cycles and to take proactive actions for maximum customer satisfaction. For making an Operational Risk Framework proactive, especially in the manufacturing domain, data-driven AI-enabled actionable insights are as vital as oil to an engine.

So there will always be operational risks in the business but the ask is can you leverage technology to identify those and act before it results in losses? Like it was said before, what can go wrong will go wrong - can YOU identify potential problems earlier?