Over the past decade, the use of artificial intelligence in supply chains has increased dramatically. AI is being used to improve the efficiency of overall operations, reduce costs and increase customer satisfaction. But the technology is very data intensive, and for it to work there must be targeted use cases that are focused on critical business workflows — not to mention systems for ingesting data that ensure the quality and integrity of information.

One of the biggest areas where AI has been successful is through the identification and prioritization of activities related to inventory optimization and shortage management, then linking them to automated workflows that increase supply chain resilience and responsiveness. These advanced technologies have shown dramatic success in breaking down the silos of decision-making between production planners, material buyers and suppliers. The result is better collaboration between these teams around the priorities that matter, and automation of those decisions where AI can accurately predict the correct response and outcome. AI is at its best when it makes work easier for individuals in the trenches. Successful applications including the automation and prioritization of inventory actions and confidence scoring.

Optimizing inventory through AI is achieved by establishing the optimal order policy and target inventory sizing level, then automatically prioritizing actions that will reduce excess inventory while maximizing production readiness. The way things are purchased is based on a series of parameters, including lead time, order quantity, demand, procurement policy and safety stock.

AI can help set those levels using a “plan for every part” (PFEP) analytical approach that draws on historical data, future demand, supplier performance and more. For example, if you run a statistical analysis of safety stock, you can go back in time and see when you had too little or too much. This enables an algorithmic approach, combined with machine learning, to inventory sizing that’s often not possible with traditional tools and spreadsheets.

Confidence scoring of procurement actions looks at supplier on-time delivery, lead times, past recommendations and other critical supply chain performance data. It evaluates each action and assesses the probability that it will be successful. The problem reported by many supply chain operations teams is that they’re already getting a lot of information, but struggling with prioritizing the critical actions that will have the biggest improvement on operational performance. Confidence scoring cuts through the noise to empower teams to make the optimal business decision.

One very successful application of AI in supply chains is to make smart recommendations — for example, using it to optimize working capital and predict future shortages. The key is that it can be used for predictive actions that will improve outcomes in both working capital and on-time delivery performance in manufacturing. Existing enterprise systems are built around handling transactions and showing you how you did in the past. AI predicts future issues accurately and prescribes specific actions, allowing systems to take some of the burden of decision-making from people, so that they can focus on more complex issues where direct intervention is required. Once you start using AI for recommendations linked with machine learning for confidence scoring, your path to automation becomes clear, and success rates are high.

The quality data is a key predicter of success. Big data and cloud-computing technologies are enabling companies to quickly connect with existing enterprise systems and process large amounts of data more efficiently. In the process, they can perform for more accurate and sophisticated predictive, prescriptive and collaborative analytics. This helps to improve supply chain decision-making and reduce waste.

Manufacturing companies are striving to get better at planning, and they need the underlying detail and data structure to make key decisions in procurement. With large amounts of historical data, AI can help companies develop an understanding of future patterns. This is the foundational piece of the technology that’s often overlooked. Data you can trust over large periods of time helps you understand how you did in the past, and predict actions that will improve the future.

Complexity in the supply chain has increased exponentially over the last decade, and decision-making and data are moving faster than ever. AI technology is transforming the supply chain industry by improving the efficiency of operations, reducing costs and increasing customer satisfaction. The goal now for manufacturers is to find a way to automate the 60% to 70% of work that is predictive and prescriptive, by relying on advanced analytics, machine learning and purpose-built workflows. Supply chain leaders who embrace these technologies can gain a competitive advantage, and position themselves for success in the rapidly evolving supply chain industry.

Richard Lebovitz is chief executive officer of LeanDNA.