Taking AI to the Next Level for Supply Chains

ISTOCK.COM/CHAYADA JEERATHEEPATANONT

While most are generally aware of the potential benefits of AI for their organizations, the value of specific applications of AI for supply chains is less clear. Recent research by Accenture found that companies using generative AI, advanced machine learning and other evolving technology for autonomous decision‑making and advanced simulations can adapt to changes more readily and adopt continuously emerging technologies more easily.

Additionally, a study by Nucleus Research found the areas of most interest in AI use in supply chains were inventory management (45%), demand forecasting (40%) and route planning (35%).²

Inventory Management

Inventory readiness requires knowing where inventory is located, including which warehouse and what position within the warehouse. AI tools can help position inventory to optimize operations based on forecasted demand, said Keith Moore, chief executive officer of AutoScheduler.AI.

Companies with multiple distribution centers can lower transportation costs and better meet customer needs with inventory located closest to customers. “Traditionally, inventory is taken by people on lifts manually checking product and displaying the data on spreadsheets,” said Sean Mitchell, vice president of customer success at MHI member Gather AI.

Because the process is cumbersome and time‑consuming, many facilities may only conduct a full inventory one, two or four times each year, he explained. “AI can interpret the image of a location, read barcodes, determine how full the location is, estimate numbers of cases and determine pallet type and depth of storage,” he said. “Our customers use a drone with a camera to scan all the racks to gather information, which collects data much faster. In fact, an operator with a drone can scan 300 to 900 locations in the time a person on a lift can scan 40 to 60 locations.”

Gather AI’s software leverages computer vision and machine learning to compare data to the warehouse management system records and identifies exceptions, which are forwarded to the inventory control team to manage the solution, said Mitchell. “Because the time required to scan inventory is reduced, we have customers who are checking inventory more frequently, which gives them more accurate information.”

Demand Forecasting

Use cases for AI‑supported planning include AI‑powered demand forecasting, scenario modeling based on external events, near‑term demand sensing and end‑to‑end delivery visibility and exception management.³

A challenge for demand forecasting or any planning activity is siloed data. When data within one facility or one area of the company is not available, the accuracy of demand forecasts declines.

“Using an AI solution that connects all distribution, manufacturing or warehouse facilities as well as all departments in an organization creates better forecasts because information can be pulled from multiple sources, not just historical data,” said Moore. “The AI algorithms create forecasts based on weather patterns, sales promotions, current inventory and historical data.” In the consumer product goods arena, more accurate forecasts support better decisions to ensure the right products are in the right place at the right time.

“With intelligent scoring, AI helps inventory planners apply economic prioritization that focuses attention on the most important opportunities, ultimately reducing bloated inventories while increasing service levels and avoiding deficits,” said Lisa Henriott, senior vice president of product marketing for Logility. “One client increased their forecast accuracy by 27%, which in turn improved inventory performance and freed up the team to focus on other strategic initiatives.”

“By leveraging an AI‑first approach, companies can dramatically improve planner efficiencies—reducing the time required to create the plan by 30%,” said Henriott. Not only does that free employees to tackle other value‑added responsibilities, but providing AI‑based tools also helps in recruiting top talent in a tight labor market.”

Route Planning

AI‑assisted route planning can be used at a local level to program robots within a warehouse or to plan delivery from the warehouse to other distribution centers to the last mile delivery. Because AI is constantly learning as more data is acquired, transparency into the entire organization produces options that lead to more efficient, cost‑effective ways to meet customers’ needs.

Dynamic lead time visibility and auto‑updates, predictive logistics disruptions, route optimization and AI chatbots for tracking are some use cases that improve transportation plans.³

The use of AI to plan transportation enables a person to look at multiple factors that affect the cost and efficiency of moving goods, said Moore. AI looks three to four steps ahead, incorporating information such as where inventory is located, what is happening at each distribution node throughout the day and how many doors or positions are available. “After evaluating the data, the recommendation might be to delay a shipment by three hours so the trailer can be filled versus shipping a partial load twice,” he said. “Automatic messages can inform customers and update other route schedulers.”