machine learning (ML) applications, and ultimately, better predictive analyses. New visibility solutions emerge regularly, says Blythe Chorn, managing director, lead, supply chain sustainability with KPMG. Yet, while many organizations would like a one-stop visibility solution, success in identifying one may come down to how companies de ne risk, which can vary by industry, geography, or other factors. For instance, a company in the food sector may source commoditized agricultural products that historically have been challenging to trace to origin. Regulations, such as the EU Deforestation Regulation (EUDR), require geospatial data on the origin of many agricultural inputs, requiring visibility back to the source, Chorn says. Yet these supply chains often are opaque. Conversely, visibility for an automotive company facing increased cybersecurity concerns likely will be focused on Tier 1 and 2 suppliers, rather than raw materials. Given these differences, organizations in this sector likely will require more customized solutions. FROM VISIBILITY TO ACTION Visibility tools that can capture data are a rst step. Technology that not only aggregates but provides insights based on the data and powered by arti cial intelligence is crucial to making supply chains more ef cient, sustainable, and adaptable to real-time demand, says Dylan Jones, senior product manager with IBM. Demand forecasting is perhaps the oldest and most proven use for AI in supply chain, Jones says. Once demand has been established, reorder levels can be optimized by incorporating other available data and applying decision management AI. That said, AI might not be able to provide all supply chain insights. For instance, the pandemic began abruptly but then rolled out over a long period of time. One might argue that there was some time to prepare and adapt to the upcoming disruption, Jones says.
“However, the full extent of the supply chain disruption that followed would have been hard to predict,” he adds. Technology platforms are increasingly using AI and machine learning for advanced predictive analytics, which is crucial for spotting potential disruptions, delays, and inef ciencies, says Sundip Naik, partner, supply chain, EY Americas logistics leader. AI systems can aggregate data from diverse sources from within and outside a company, such as ERP systems, IoT sensors, weather forecasts, and supplier performance records, to provide real- time monitoring, for example. Machine learning models can process data to detect anomalies and patterns, such as seasonal demand spikes or supplier failure risks. Through predictive modeling, these can help detect and forecast potential issues, such as delays or bottlenecks. Along with predictive capabilities, AI tools and platforms can employ simulations to assess risks, evaluate what-if scenarios, and provide actionable recommendations, such as identifying alternative suppliers, rerouting goods, or adjusting inventory. Digital twins powered by AI can create virtual supply chain replicas to simulate disruptions and test responses, while machine learning can continuously update these models, creating an evolving, insightful platform to aid
decision-making. By leveraging AI, ML, and visibility solutions, companies have the tools to analyze complex scenarios, evaluate various options, and identify actions that can optimize operations, minimize costs, and boost customer satisfaction. While data is the foundation of visibility, it must be accurate, complete, and relevant to be useful. Titze provides this example: a company implements real-time transport visibility across multiple modes but lacks a crucial link to its bills of lading. As a result, the company could track shipments but couldn’t identify which products were in each shipment. If the system had integrated shipment data with bills of lading, the company could have understood which products were affected by delays, helping them better assess impacts on production and sales. “The key,” Titze says, “is connecting different data sources to the speci c use case.” DATA ACCESSIBILITY AND ACCURACY Concerns over data quality issues shouldn’t be a reason to resist taking action, however. “We have come to a point in time where organizations can no longer be avoidant,” Jones says. Moreover, AI or rules-based processing can help to identify and address data quality problems.
Sensor technologies, such as RFID, LTE, and fusion models, enhance and facilitate visibility in the warehouse, making distribution operations more ecient and adaptable to real-time demand. Tracking data from sensors is increasingly being integrated with AI solutions.
July 2025 • Inbound Logistics 141
Powered by FlippingBook