Taming extreme supply chain disruptions using next-gen risk assessment
Supply chain risks in the coming era
Supply chain disruptions are, without a doubt, one of the most serious challenges faced by manufacturers and retailers on a day to day basis. The COVID-19 pandemic of 2020 made the wider population of consumers keenly aware of this problem as the sudden changes brought about by the lockdown filtered throughout the whole economy and led to empty shelves at the retailers. With ‘black swan’ type political events like Brexit, and steady background uptick of ‘once-in-a-lifetime’ events brought about by climate change, more extreme disruptions to supply chains are likely to become the new normal in the very near future. If not brought under control, these extreme disruptions have the potential to not just hurt the profitability/viability of companies, but lead to wider societal problems such as unemployment, lack of access to basic necessities and general social unrest.
Under these circumstances, it is absolutely necessary that the wider community of manufacturers, retailers, suppliers, as well as the insurer, legal and accounting firms supporting them, embark on a concerted industry-wide effort to correctly assess the risks in the chain as close to real time as possible, and take mitigating actions subsequently. Fortunately this is an era where advanced technologies like artificial intelligence (AI), internet of things (IoT), 5G, blockchain/distributed ledger technologies (DLT) are all coming into their own and can play a deciding role in meeting these challenges. This article will explore the gaps in the industry that make it difficult to assess the risks properly, the business value of mitigating them, and the role that these technologies can play in addressing these problems by making the supply chains more robust and adaptive. The latter is made concrete in the context of an InnovateUK funded commercial R&D project called KnowRisk being undertaken by a consortium of which Digital Catapult is a partner.
Failure to properly assess risks in the supply chain
The main reason it is difficult to properly assess risk within the supply chain is because each party in the chain lacks relevant, up to date information about the actions and circumstances of the other parties in the value chain. At a high level, the parties involved are retailers, manufacturers, suppliers, insurers, and legal and accounting firms and it is illuminating to understand this lack of information from the perspective of each party, and its consequences.
Starting at the end of the chain, the retailers lack information on the current business condition of its suppliers, including manufacturers, and any other kind of natural or political events/disasters that may be interfering with the suppliers’ ability to meet their obligations. This may result in empty shelves/lack of availability of items, waste, increased or unnecessary warehouse costs, customer churn and lost business. In extreme cases this may even lead to the shops closing. Manufacturers and suppliers face similar problems in turn, ast they have no information on the risks that their suppliers are facing and how the suppliers’ risks have increased or been mitigated. This makes it difficult for all involved to plan ahead and be proactive rather than reactive with regards to problems in the supply chain. Ultimately, this results in delays in the production line and inability to meet commitments made to the retailers.
The insurance companies who insure retailers, manufacturers and suppliers are another party in the supply chain that ,in principle, have a view of all the risks involved by virtue of their risk assessment activities. Unfortunately, these companies are often not in a position to help their clients understand their risks for a variety of reasons. Firstly, the risk assessment process is completely manual, expensive and very slow. This makes it impractical to conduct such activities more than once after a policy has begun. Indeed, it is often impossible to do even a single review of all assets, and group level statistics are used to arrive at risk scores.
Secondly, the information systems used by insurers are either not automated enough, or not standardised, so that the data exists in silos that cannot talk to each other. This means that insurers are not able to have a global view of the risk, unable to integrate external information sources that may augment their data, update current policies based on any mitigation information that comes in, or even incorporate claims information when designing their future policies. Altogether, this causes insurers to misprice their premiums, exposes them to large risks leading to avoidable losses and lost business because they cannot assess all clients.
The final party in the supply chain to be considered are legal and accounting firms that provide their services to the retailers, manufacturers and suppliers. These firms lack information about various risks in the supply chain in a manner similar to the insurance firms. As a result the contracts they create are not designed to make the business operations more efficient with timely information about which parties have succeeded or failed to meet their obligations, or whose rights have changed. Rather, contracts are created in a defensive fashion to ensure clients are protected in case anything goes wrong. Another significant issue that accounting firms face is that clients are required, by law, to switch firms every five years. However, the lack of a standardised data format makes this a very slow and expensive process — indeed the total cost of accounting operations is doubled due to this problem.
In summary, participants within a supply chain are suffering from a chronic lack of relevant information necessary to mitigate risks. As a result they are predominantly reacting to problems as they arise rather than being proactive and efficient. Under these circumstances it is no surprise that this leaves the whole chain highly vulnerable to disruption, where a single point of failure can propagate through the entire chain causing significant losses to all involved.
Accurately assessing risks in the supply chain: The KnowRisk project
This section will discuss some of the issues that need to be addressed to fix the information gaps identified in the previous section and how they are being addressed in the context of the KnowRisk project. KnowRisk is an InnovateUK funded commercial research and development project that aims to serve as a proof of concept for a modern supply chain that utilises advanced technologies such as AI, DLT and geospatial intelligence (GEOINT) to help assess and mitigate risks in a timely manner. The project will also serve as a first step toward building out such a modern system at scale and rolling that out. The partner companies in the KnowRisk project are SweetBridge, Engine B, Cystellar, Digital Catapult, Intelligent AI and Industria, with Sweetbridge the leading partner.
Below is a list of some of the problems that need to be fixed to modernise and properly risk assess the supply chain, and how that is being done within Knowrisk by the relevant partner(s).
Data standardisation: A standardised set of protocols and data formats needs to be created to let all the parties communicate information relevant to the supply chain quickly and accurately. Engine B is a consortium of nine accounting firms and is leading the creation of a standard set of data formats for exchange of accounting data. It is estimated that this will reduce the cost by half, and enable additional automation. Sweetbrdige is also integrating this data with additional information from the corporate clients and storing it in a standardised way.
Data integrity. The data being shared by the various participants in the supply chain is confidential and as such the data has to be stored and shared using a trusted process. To address this, Industria is using DLT to store the data from all the parties to ensure that the data is being stored in a distributed, transparent, secure manner, whilst remaining private.
Automated contracting. To understand the current state of the risk profile of the supply chain it is very important to ensure that all the contracts involved are up-to-date in terms of which obligations have been fulfilled, which have not, which ones have been breached and what impact that has had on the rights and obligations of others. Sweetbridge will be using its proprietary technology to collect information from a variety of businesses and insurers to build a proof of concept automatic contracting solution.
Automated risk assessment. To have accurate and timely supply chain risk assessments, it is important to have fast and reliable methods for calculating risk based on the most recent information about the various parties involved. Within KnowRisk, a component will be created to convert natural language PDF risk reports into computer readable plain text by Intelligent AI, then a machine learning (ML) system created by Digital Catapult will take the reports and convert them into risk mitigations and actual risk scores. Intelligent AI will additionally focus on integrating information from the reports and the other sources mentioned here and presenting them in a dashboard to give a coherent and comprehensive view of risks involved.
Information sharing for risk assessment. Machine learning methods benefit greatly from being able to combine data owned by different participants (such as insurers and businesses amongst others), where the data is confidential and generally cannot be shared, let alone combined. To counter this, the ML system will be embedded within a federated learning platform, developed at Digital Catapult, which will allow the system to use all the data, whilst ensuring data confidentiality, without ever copying or moving the data from the data-owners systems. Crucially, this automated system will serve as a prototype for calculating risk scores, not just from risk reports, but from a variety of potentially private sources and a variety of modalities as soon as new information becomes available about the relevant component in the supply chain. The information here will feed into the automated contracting solution to provide a even more accurate view of the current state of rights and obligations
Integrate third party data source. To get a complete picture of the risks involved it will be important to integrate relevant information from third party sources. As a proof of this idea, Cystellar will further enrich the overall view of the risks involved by using GEOINT technologies. Cystellar uses ML techniques on geospatial data from satellites to calculate risks that may result from hazards such as flooding and wildfires.
Conclusion
We are currently observing supply chain disruptions that are much more severe due to the world becoming fundamentally more complex and interdependent, as well as the impacts seen due to climate change. This is likely to be the new normal, and there is an urgent need to upgrade and overhaul how supply chain risks are assessed and mitigated. On the other hand, this is also a time where technologies such as AI, DLT, GEOINT are coming into their own in the business world and KnowRisk project both aims to show how these may be used to tame the supply chain disruptions and also serve as a first step toward modern, adaptive and robust supply chains.
Acknowledgements
The article includes contributions from Ben Ramsden, Partnership Manager at Digital Catapult for Food and Drink.
