Updated: Sep 3, 2019
Credit by: George Lawrie
It’s now just over 12 months since we first published research on digital supply chains. Some Forrester clients have made progress with integrated business planning and with control towers. And they are on the lookout for blockchain opportunities. But the most common question we hear is about the role of AI in supply chains. The major applications of AI in supply chains are in demand management, warehouse execution, and chatbots serving procurement professionals.
1. Demand Management
For decades, forecasts have adapted to changing patterns of consumption. But B2B and B2C purchasers and shoppers now have more choice than ever. Research chemists scour the web for innovative new ingredients for their formulation trials. Grocery shoppers substitute fresh for frozen or packaged food. Companies have to reinvent demand management. Sophisticated machine-learning solutions, from specialists such as RELEX, help retailers like Ahold Delhaize include variables such as weather in better forecasts that reduce waste in fresh food.
IBM applied cognitive technologies to its own supply chain history, preference, and best practices. It developed “Resolution Rooms” with best practices to respond to supply chain disruption. IBM’s Resolution Rooms helped to mitigate an industrywide flash shortage and avoided a potential revenue impact of $60 million or more due to late customer order shipments. In 2016, IBM recorded a 52% reduction in expediting costs.
2. Warehouse Execution Systems
Automated guided vehicles (AGVs) have been features of warehouses and factories for decades but have only recently become autonomous. Their autonomy is in part the result of superior computer vision. Computer vision exploits more powerful chips and cheaper memory to enable local recognition of objects and obstacles. But visual computing has also enabled companies to deploy three AI applications with significant impact for digital operations: load analytics, pick to light, and goods to person.
Visual computing solutions such as Zebra’s SmartPack use AI to ensure that warehouse workers load trailers in the best way to use available space and to protect cargo.
Pick To Light
PICK TO LIGHT, part of AIOI Systems (Japan), uses AI to help companies like Nespresso accurately pick and ship high volumes of customer orders.
Pick to light
Goods To Person
Traditional warehouse management systems all include some kind of pick route optimization. But modern material handling moves the goods and, in some cases, the whole pick face to the person. This boosts fulfillment responsiveness to close the gap with impatient customers’ expectations.
Software vendors like Deposco and Manhattan Associates long ago merged commerce and warehouse management applications to slash order time for companies such as Central Vacuum Stores and Forever Direct, respectively.
They use sophisticated analytics to fulfill orders from the best locations, taking into account order urgency and landed cost elements such as duty or freight cost. Modern retail platforms in China use AI to compress order cycles to as little as 30 minutes from order entry to fulfillment.
Companies such as Swisslog help improve efficiencies autonomously by learning from data such as customer ordering behavior on how best to “flow” work through the warehouse. Swisslog’s predictive analytics can even anticipate the warehouse operations impact of eCommerce marketing campaigns or changes in weather.
Goods to person
3. Chatbots In Procurement
Of course, there are chatbots that can help in tracking orders and deliveries. UPS bots, for example, can talk to customers via Google Assistant using devices like Google Home. But certified SAP software partner Unvired deploys its Chyme bot platform to enable procurement to work more effectively. It generates purchase orders based on AI-enabled chat with enterprise systems such as configure price quote (CPQ) or service management.