A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by the Construction Tech Review Advisory Board.
Ernst & Young Advisory Pte. Ltd.
Building a supply chain that learns to think
Manik Bhandari
The application of AI technologies to supply chains, particularly cognitive computing technologies, has also been seen to deliver compelling business returns. For example, in recent years, we have seen organizations achieving 2-3 percent uplift in sales and market share, 15-20 percent reduction in working capital, 8-12 percent reduction in cost of goods sold and 5-6 percent increase in operating margins as a result of using AI.
Understanding cognitive computing
Many organizations may not understand what AI is, let alone the function of a subset technology like cognitive computing.
The variety of AI technologies is as wide-ranging as the human mind’s capabilities. AI technology is being used in business intelligence tools to process data and interpret data, or in robotic process automation tools to automate routine rule-based decisions, or in the case of cognitive computing, to interpret data to make judgment-based decisions and evaluate its own decision-making efficacy. The latter capability to learn and discern between good and bad decisions based on the data it collects and analyzes, is what sets cognitive computing apart from other AI technologies.
Equipping supply chains with this dynamic learning ability is proving to be a powerful differentiator in an ever-changing geo-political business landscape, where consumer trends are constantly shifting, unpredictable climate conditions are impacting resource availability, and trade policy developments are changing distribution routes.
Cognitive computing technologies are able to interpret these changes and help the supply chain adapt to, sense and shape demand; synchronize supply with market needs; optimize the distribution of goods and equipment across the value chain; predict and reduce supply chain uncertainty; and institutionalize this knowledge in the system.
Making it a success
While the business value of cognitive computing technology is clear, many organizations struggle to implement it in their supply chain. Organizations need to understand that unlike traditional IT solutions, developing cognitive computing solutions requires a different approach.
Firstly, cognitive computing solutions are trained and not coded, meaning that the model needs to be taught by human trainers who will rate the cognitive system’s quality of answers. Most organizations make the mistake of underestimating the resources required for training. As the model learns, it will also evolve, likely resulting with a different set of goals and requirements from the original framework. This requires organizations to embrace an iterative approach, instead of expecting the system to deliver results from the start.
Secondly, in view of the resource-intensive demands of these solutions, organizations should identify the most critical supply chain pain points and articulate how it is linked to business goals, or risk conducting merely theoretical exercises.
Nailing the right problem is just the first step. Many organizations forget to predefine their target KPIs. Without a clear benchmark and target, organizations are at risk of overtraining their systems and wasting resources.
Successfully transforming supply chain management with cognitive computing will require a business-driven strategy, working with the ecosystem of solutions, and evolving the workforce
Upon conducting an assessment of their data, organizations may discover that their data quality falls below their expectations. Yet, this should not become a roadblock. For example, in helping an organization to predict its vessel timings, we worked with our client to leverage a common and widely available data source – satellite information, which required little commercial investment.
Fourthly, the plethora of literature, software, applications, and use cases on cognitive computing could be overwhelming to supply chain executives. Just as organizations do not require perfect data sets to embark on their cognitive transformation journey, neither do they need the perfect algorithm. Instead of trying to understand this complex technology on their own, organizations can plug into the cognitive computing ecosystem to build close business partnerships that can come together to experiment with emerging methodologies.
There are rarely one-size-fit-all solutions in tackling supply chain challenges with artificial intelligence and a right portfolio of solutions will be necessary to provide an end-to-end solution. The ecosystem approach is one of the proven methods to address this issue.
Organizations also need to recognize the emerging cognitive supply chain specialist skillsets required and shift their workforce accordingly. The workforce will have to evolve as supply chains are embedded with advanced analytics, cognitive and artificial intelligence solutions.
The successful cognitive transformation of the supply chain is not merely an IT project, requiring a mix of both suits and jeans. As such, organizations will need to equip their supply chain professionals to be both analytics-savvy as well as operations-focused to be able to translate a supply chain challenge into an analytical problem statement, and work with technologists to develop cognitive computing solutions that have a tangible impact on the realities of running daily operations.
Securing a competitive edge
AI and cognitive computing are likely to stay for the long term and disrupt how organizations run their supply chains now and into the future. For the organizations that embark on cognitive transformation with the end business goal in mind, experiment with the ecosystem of solutions, and transform their workforce based on the skills of today and into the future, these emerging technologies provide a great opportunity to create sustainable competitive advantage.
