Companies may be underestimating the potential of artificial intelligence (AI) in tackling consumer demand volatility, judging by their delay in committing serious investment funding to the technology, according to a study by Relex Solutions, a Finnish supply chain software provider.
The study showed that 57% of retailers and consumer packaged goods (CPG) companies plan to invest in predictive and generative AI in the next 3-5 years. But they may not invest very much; a result from the same study also found that AI and machine learning (ML) ranked just fifth in importance for overall technology spend, coming in behind: enhancing e-commerce capabilities, improving inventory management, demand forecasting, and leveraging data analytics.
Despite their hesitancy to commit to AI as the solution, respondents were clear about their need to improve demand forecasting accuracy, Relex Solutions said. Companies ranked the biggest threats to improving supply chain efficiency and accuracy over the next three years as: rapidly shifting consumer demand volatility (55%), global events & disruptions (50%), and inaccurate sense of customer-specific demand (43%). And a whopping 94% of respondents also said they have been impacted by social media influencing or "de-influencing" over the last 12-24 months – another sign that anticipating and sensing demand surges across all channels will be especially critical to success, Relex Solutions said.
The statistics come from the company’s “State of Supply Chain 2024: Retail and CPG Dynamics” report, conducted by Researchscape, which surveyed 285 retail, CPG, and wholesale leaders globally in February 2024.
"The retail and CPG industries continue to face complex, global challenges that require actionable insights to accurately predict, anticipate, and manage consumer demand,” Laurence Brenig-Jones, VP of Strategy & Marketing at Relex Solutions, said in a release. “To thrive in this new reality, companies must fundamentally transform their approach to supply chain management by breaking down silos, embracing new technologies like AI and ML, and fostering a culture of collaboration and agility.”
In other results, the survey showed that:
The top three most essential capabilities for retailers to manage consumer demand and inventory levels are real-time inventory visibility (45%), customer demand sensing (45%), and inventory optimization tools (43%)
More than half (53%) are expanding supplier base / sourcing options to add sourcing redundancy to diversify and mitigate risks associated with disruptions
To navigate macro-economic factors like inflation, CPGs are focused on adjusting inventory and production strategies via monitoring demand signals (48%), building in safety stock or keeping higher inventory levels (22%), and inventory turnover as part of their inventory and production strategies
Robots are revolutionizing factories, warehouses, and distribution centers (DCs) around the world, thanks largely to heavy investments in the technology between 2019 and 2021. And although investment has slowed since then, the long-term outlook calls for steady growth over the next four years. According to data from research and consulting firm Interact Analysis, revenues from shipments of industrial robots are forecast to grow nearly 4% per year, on average, between 2024 and 2028 (see Exhibit 1).
EXHIBIT 1: Market forecast for industrial robots - revenuesInteract Analysis
Material handling is among the top applications for all those robots, accounting for one-third of overall robot market revenues in 2023, according to the research. That puts warehouses and DCs on the cutting edge of robotic innovation, with projects that are helping companies reduce costs, optimize labor, and improve productivity throughout their facilities. Here’s a look at two recent projects that demonstrate the kinds of gains companies have achieved by investing in robotic equipment.
FASTER, MORE ACCURATE CYCLE COUNTS
When leaders at MSI Surfaces wanted to get a better handle on their vast inventory of flooring, countertops, tile, and hardscape materials, they turned to warehouse inventory drone provider Corvus Robotics. The seven-year-old company offers a warehouse drone system, called Corvus One, that can be installed and deployed quickly—in what MSI leaders describe as a “plug and play” process. Corvus Robotics’ drones are fully autonomous—they require no external infrastructure, such as beacons or stickers for positioning and navigation, and no human operators. Essentially, all you need is the drone and a landing pad, and you’re in business.
The drones use computer vision and generative AI (artificial intelligence) to “understand” their environment, flying autonomously in both very narrow aisles—passageways as narrow as 50 inches—and in very wide aisles. The Corvus One system relies on obstacle detection to operate safely in warehouses and uses barcode scanning technology to count inventory; the advanced system can read any barcode symbol in any orientation placed anywhere on the front of a carton or pallet.
The system was the perfect answer to the inventory challenges MSI was facing. Its annual physical inventory counts required two to four dedicated warehouse associates, who would manually scan inventory to determine the amount of stock on hand. The process was both time-consuming and error-prone, and often led to inaccuracies. And it created a chain reaction of issues and problems. Fulfillment speed is one example: Lost or misplaced inventory would delay customer deliveries, resulting in dissatisfaction, returns, and unmet expectations. Productivity was also an issue: Workers were often pulled from fulfillment tasks to locate material, slowing overall operations.
MSI Surfaces began using the Corvus One system in 2021, deploying a small number of drones for daily inventory counts at its 300,000-square-foot distribution center (DC) in Orange, California. It quickly scaled up, adding more drones in Orange and expanding the system to three other DCs: in Houston; Savannah, Georgia; and Edison, New Jersey. The company plans to add more drones to the existing sites and expand the system to some of its smaller DCs as well, according to Corvus Robotics spokesperson Andrew Burer.
Those expansion plans are based on solid results: MSI’s inventory accuracy was about 80% prior to the drone implementation, but it quickly jumped to the high 90s—ultimately reaching 99%—after the company initiated the daily drone counts, according to Burer.
“We actually had an incident early on where one of the forklift drivers ran into the landing pad, rendering it inoperable for about a week while the Corvus team fixed it,” Burer recalls. “When we restarted the system, we noticed MSI’s inventory accuracy had dropped down to the 80s. But after flights resumed, accuracy quickly improved back to near perfect.” He adds that such collisions are rare as Corvus mounts landing pads high off the floor to avoid impacts but that accidents can still happen.
Overall, the system has helped speed warehouse operations in two key ways: First, the accuracy improvement means that associates no longer waste time searching for missing material in the warehouse. And second, the associates who used to conduct the physical inventory counts have been reallocated to picking and replenishment—creating a more efficient, and optimized, workforce.
A SAFER, MORE EFFICIENT WAREHOUSE
Robot maker Boston Dynamics is well-known for its Stretch and Spot industrial robots, both of which are at work in warehouses and DCs around the world. Earlier this year, Stretch made its debut in Europe, teaming up with Spot at a fulfillment center run by German retail company Otto Group. The deployment marks the first time Stretch and Spot are being used together—in a partnership designed to improve Otto Group’s warehousing operations by increasing efficiency and making warehouse work safer and more attractive to workers.
The partnership is part of a two-year project in which Boston Dynamics will deploy dozens of its warehouse robots in Otto Group’s European DCs. The first location is a fulfillment site operated by Hermes, the company’s parcel delivery subsidiary, in Haldensleben, Germany—a facility that handles as many as 40,000 cartons of goods on peak days.
At the site, Stretch—which is a mobile case-handling robot—autonomously unloads ocean containers and trailers, using its advanced perception system to pick and place boxes onto a telescoping conveyor inside the container or trailer. Spot—a quadruped robot—helps with predictive maintenance by collecting thermal data and performing acoustic and visual detection tasks throughout the facility to reduce unplanned downtime and energy costs. One of Spot’s jobs is to detect air leaks in the facility’s warehouse automation systems; future duties may include conveyor vibration detection, according to leaders at Otto Group.
Both Stretch and Spot will help the Haldensleben facility run more efficiently, especially during fall peak season when volume increases and work intensifies. The addition of Stretch addresses safety and comfort issues as well: Trailer unloading—a process that entails repeatedly lifting and moving heavy boxes inside a trailer, which can be dark, dirty, cold, and/or hot, depending on the weather—tends to be unappealing to workers. Along with reducing the amount of labor required, automating these tasks will have the added benefit for European facilities of helping them comply with EU (European Union) regulations limiting the amount of time workers can spend in those conditions.
Essentially, the robots are making life easier on the warehouse floor and for the company at large.
“Stretch is going to have a ton of benefits for customers here in the EU,” Andrew Brueckner, of Boston Dynamics, said in a recent case study on the project.
The trucking industry faces a range of challenges these days, particularly when it comes to load planning—a resource-intensive task that often results in suboptimal decisions, unnecessary empty miles, late deliveries, and inefficient asset utilization. What’s more, delays in decision-making due to a lack of real-time insights can hinder operational efficiency, making cost management a constant struggle.
Truckload carrier Paper Transport Inc. (PTI) experienced this firsthand when the company sought to expand its over the-road (OTR), intermodal, and brokerage offerings to include dedicated fleet services for high-volume shippers—adding a layer of complexity to the business. The additional personnel required for such a move would be extremely costly, leading PTI to investigate technology solutions that could help close the gap.
Enter Freight Science and its intelligent decision-recommendation and automation platform.
PTI implemented Freight Science’s artificial intelligence (AI)-driven load planning optimization solution earlier this year, giving the carrier a high-tech advantage as it launched the new service.
“As PTI tried to diversify … we found that we needed a technological solution that would allow us to process [information] faster,” explains Jared Stedl, chief commercial officer for PTI, emphasizing the high volume of outbound shipments and unique freight characteristics of its targeted dedicated-fleet customers.
The Freight Science platform allowed PTI to apply its signature high-quality service to those needs, all while handling the daily challenges of managing drivers and navigating route disruptions.
STREAMLINING PROCESSES
Dedicated fleets face challenges that evolve from day to day and minute to minute, including truck breakdowns, drivers calling in sick, and rescheduled appointment times. PTI needed a tool that allowed for a real-time view of the fleet, ultimately enabling its team to adjust truck and driver allocation to meet those challenges.
The Freight Science solution filled the bill. The platform uses advanced analytics and algorithms to give carriers better visibility into operations while automating the decision-making process. By combining streaming data, a carrier’s transportation management system (TMS), machine learning, and decision science, the solution allows carriers to deploy their fleets more efficiently while accurately forecasting future needs, according to Freight Science.
In PTI’s case, Freight Science’s software integrates with the carrier’s TMS, real-time electronic logging device (ELD) data, and other external data, feeding an AI model that generates an optimized load plan for the planner.
“We’re an integrated data analytics company for trucking companies,” explains Matt Foster, Freight Science’s president and CEO. “We’re talking about AI.”
The benefits of the real-time data are difficult to overstate.
“We’ve been able to execute in the toughest of situations because we’ve got real, live data on how long each event is actually going to take and a system to aid and even automate the decision-making process,” says Chad Borley, PTI’s operations manager. “From what traffic patterns we are battling in the morning and evening with rush hour and things like that, to the impact of additional miles to a route, or even location-specific dwell times, it’s been a huge differentiator for us.”
REALIZING RESULTS
A case in point: the collapse of Baltimore’s Francis Scott Key Bridge in March. PTI was scheduled to go live with a new dedicated account in the area just days after the collapse, which would mean rerouting and the potential for longer transit times. Instead of recalculating based on assumptions or latent data, PTI was able to reroute freight based on real-time information and analytics to give the customer timely updates.
“With the bridge going out, that changed our ability to make as many turns a day as the customer would expect,” Stedl explains. “But one of the things Freight Science could do [was to] quickly [assess] how much of an impact that traffic would have [and] what the turns [would] be based on what’s happening on the ground.
“So we were able to go back to the customer and readjust expectations in a real way that made sense, using data. Now expectations can be reset¾we’re not asking for forgiveness when there’s no reason for it.”
The system’s advanced algorithms make load planning more cost-effective and scalable as well. The platform allows PTI to monitor trucks, trailers, and driver hours in real time, recommending additional loads with remaining driver hours that would otherwise be wasted.
And they’re doing it all with much less. Stedl says tasks that used to require five people and hours of work can now be accomplished by one person in mere minutes, improving productivity and profitability while reducing labor and operational costs.
Terms of the deal were not disclosed, but Aptean said the move will add new capabilities to its warehouse management and supply chain management offerings for manufacturers, wholesalers, distributors, retailers, and 3PLs. Aptean currently provides enterprise resource planning (ERP), transportation management systems (TMS), and product lifecycle management (PLM) platforms.
Founded in 1980 and headquartered in Durham, U.K., Indigo Software provides software designed for mid-market organizations, giving users real-time visibility and management from the initial receipt of stock all the way through to final dispatch of the finished product. That enables organizations to optimize an array of warehouse operations including receiving, storage, picking, packing, and shipping, the firm says.
Specific sectors served by Indigo Software include the food and beverage, fashion and apparel, fast moving consumer goods, automotive, manufacturing, 3PL, chemicals, and wholesale / distribution verticals.
Schneider says its FreightPower platform now offers owner-operators significantly more access to Schneider’s range of freight options. That can help drivers to generate revenue and strengthen their business through: increased access to freight, high drop and hook rates of over 95% of loads, and a trip planning feature that calculates road miles.
“Collaborating with owner-operators is an important component in the success of our business and the reliable service we can provide customers, which is why the network has grown tremendously in the last 25 years,” Schneider Senior Vice President and General Manager of Truckload and Mexico John Bozec said in a release. "We want to invest in tools that support owner-operators in running and growing their businesses. With Schneider FreightPower, they gain access to better load management, increasing their productivity and revenue potential.”
Terms of the acquisition were not disclosed, but Mode Global said it will now assume Jillamy's comprehensive logistics and freight management solutions, while Jillamy's warehousing, packaging and fulfillment services remain unchanged. Under the agreement, Mode Global will gain more than 200 employees and add facilities in Pennsylvania, Arizona, Florida, Texas, Illinois, South Carolina, Maryland, and Ontario to its existing national footprint.
Chalfont, Pennsylvania-based Jillamy calls itself a 3PL provider with expertise in international freight, intermodal, less than truckload (LTL), consolidation, over the road truckload, partials, expedited, and air freight.
"We are excited to welcome the Jillamy freight team into the Mode Global family," Lance Malesh, Mode’s president and CEO, said in a release. "This acquisition represents a significant step forward in our growth strategy and aligns perfectly with Mode's strategic vision to expand our footprint, ensuring we remain at the forefront of the logistics industry. Joining forces with Jillamy enhances our service portfolio and provides our clients with more comprehensive and efficient logistics solutions."