AI in Supply Chain

As the supply chain sucker-punched its way into a Hollywood starring role during the pandemic, the consumer goods industry was forced to reckon with a confluence of new, untested challenges. Not only did empty shelves strain retail relationships and consumer brand perceptions, but persistent labor challenges served as perpetual pours of gasoline on the multi-year fire. 

While the pandemic is fading into the rearview mirror and catalog of bad dreams, supply chain complexity remains. Port congestion, political strife, and labor difficulties continue, while mounting sustainability and ethical labor mandates require manufacturers and retailers to have more visibility into their supply chains to maintain compliance. After spending much of its life garnering a blank stare at best, companies are now leveraging technology to not only bolster the resiliency of their supply chain, but to even transform it into a competitive advantage. In fact, just a scant 10% of companies believe that supply chain technology is not a source of competitive advantage, according to Gartner research shared with CGT.     

“Supply chain became the ‘cool kid’ in the last three years, but there’s a lot to live up to because nowadays supply chain is both a dinner table and a boardroom conversation,” noted Ashok Viswanathan, Best Buy director of supply chain analytics, at CGT’s Analytics Unite event. 

Artificial intelligence is one such technology that holds the potential to strengthen every link in the supply chain. Nearly every aspect of the supply chain will be impacted by robotics and AI by 2030, according to Coresight Research, driving significant benefits in warehouses and the last mile. 

In Unpacked, CGT's new series of Tech Explainers, we break down how AI is being used in the supply chain, its benefits and drawbacks, and where today's leading experts see the future potential. 

What is the role of AI in logistics and supply chain?

Artificial intelligence can play a role throughout the entire consumer goods supply chain, from demand forecasting to last-mile delivery to product returns to waste management. Thanks to its ability to gather, organize, and analyze data at lightning-fast speeds — speeds that are simply impossible for humans — AI can help supply chain and logistics leaders make decisions more quickly and accurately.   

The technology also has the ability to help companies proactively identify potential challenges so they can minimize them or avoid them completely, and it can identify optimization opportunities that would have otherwise gone undetected. Gartner even cited “actionable AI” as one of its top supply chain technology trends for 2023.   

“AI is enabling businesses to have added intelligence and incremental predictions into their supply chains,” says Manish Ghosh, Blue Yonder’s corporate VP, industry strategy, consumer industries. “While planning and forecasting were instrumental prior to the pandemic, recent developments in AI have dramatically enhanced these capabilities. Companies are leveraging AI to predict item returns, supplier reliability, consumer price elasticity, labor needs, and much more.” 

As for generative AI in the supply chain, one of the most buzzed-about subsets of artificial intelligence? While saddled with challenges, it’s nonetheless expected by many to serve as a trusted, intelligent co-pilot companies will be able to leverage to augment and support their decision-making and problem-solving. 

How can AI be applied to supply chain activities? 

AI can be applied to supply chain activities in a variety of methods. When consumer goods manufacturers were surveyed in the CGT/RIS News Retail & Consumer Goods Analytics Study about the top three ways they’re leveraging AI/ML, supply chain activities crowded the top of the list. 
 

Here are just some of the specific ways AI can be used in the supply chain: 

Demand planning and demand forecasting: AI can improve demand planning and forecasting by quickly gathering, organizing, and analyzing data from both internal and external sources so manufacturers have a deeper understanding of the products they should make. Rather than relying on their traditional methods of examining historical data, AI can analyze weather patterns, consumer behavior trends, social media activity, and more to provide granular recommendations.  

Inventory management: AI can help companies identify and optimize the ideal locations to ship, store, and sell their products and services. Ready access to consumer behavior data and its impact on selling trends could help brands manage through seasonal demand and increase sales through hyper-localized assortment strategies. 

Customer service: Think of it as Microsoft’s Clippy on steroids: Companies can use AI-infused chatbots to efficiently and intelligently manage common complaints and issues from their consumer and retail customers. 

Sustainability: Companies can use AI to help advance their sustainability goals and stay compliant with regulatory mandates, such as by optimizing their energy use in factories and/or their transportation capacity to decrease emissions. The technology can also be leveraged to analyze the most sustainable and ethical location and method to source their raw materials. One AI startup is even leveraging the technology in the form of cameras that track and analyze waste objects, developing a shareable map waste managers can use to improve operations efficiency. 

Cross-functional decision-making: Artificial intelligence can be used to bolster a company’s decision-making capabilities and improve collaboration.  

Workforce optimization: By using Ai within their employee operations, companies can optimize labor scheduling to improve employee efficiency, such as identifying the ideal number of employees in a particular location. Using artificial intelligence in this use case enables them to potentially benefit employee safety and satisfaction. 

How does Artificial Intelligence improve the supply chain? 

At its core, AI can improve supply chain management and execution by increasing efficiency, reducing costs, and advancing a safer and more collaborative supply chain. It can help companies advance their ESG goals by increasing visibility into ethical and sustainable practices, as well as enable them to take proactive measures to decrease energy usage and harmful emissions.    

What are some disadvantages of artificial intelligence? 

There is no perfect technology; as such, there are a number of challenges of AI in the supply chain that leaders must weigh as they decide whether to invest. In an Equinix survey of 2,900 IT decision-makers in 29 markets, 42% of IT leaders said they weren’t very comfortable that their infrastructure was capable of accommodating AI, while 41% doubted their team's ability to implement the technology.

Among the obstacles to leveraging artificial intelligence and machine learning:   

Data requirements: Data can be a significant barrier when leveraging AI, especially with regards to machine learning, according to Amber Salley, senior director analyst at Gartner. A very large volume of data is required in order to have the necessary number of observations to identify patterns with machine learning, and many organizations may simply not have the volume of data needed at the right level of granularity to receive value out of the AI outputs.  

Skill set: Bells and whistles are just that if you don’t know what you’re doing — or if you’re too stuck in legacy habits. “Another big challenge is the skill sets of the planners or the resources themselves and getting them to rethink what the supply chain will be with the use of AI to enable more automation,” says Salley. “That is also included with identifying what the skill sets of the supply chain professionals need to be and how you might need to upskill the skill sets of professionals is a big one.” 

Without the right training, companies might not not know what to do with the output that they have, she notes. 

Cost: As with all emerging technologies, there can be a significant cost to leveraging AI within operations, and companies may have difficulty proving the ROI to their key stakeholders. For Tyson Foods, the ROI for its AI-infused supply chain control towers can be more obvious in cases where capabilities lead directly to measurable improvement such as gains in demand forecast or minimizing distressed inventory, according to Adam Clark, VP of IT business shared service at Tyson Foods. “More common are the general benefits of speed to decision and better-informed decision-making,” which he agrees can be difficult to initially quantify and track. 

Bias and security: The increase in AI adoption is also spurring more conversation surrounding the importance of leveraging it ethically. AI, just like humans, can carry biases, and it’s also accompanied by a host of privacy concerns. Generative AI, meanwhile, has become well known for its ability to confidently supply false information. 

As such, developing an ethical and responsible AI policy is a critical component of any AI strategy, and companies like PepsiCo are collaborating externally with such organizations as the Stanford Institute for Human-Centered Artificial Intelligence. 

“At the very top, before you start writing a line of code, [you have] to activate a trustworthy AI environment,” says Reggie Townsend, VP of data ethics at software provider SAS. “You've got to start with some measure of oversight. You’ve got to think about what your operation is going to look like. You've got to make sure you've got adequate performance and risk mitigation in place, and you've got to work on building a culture that is ethical by design.”

To do this, Townsend offers these steps as a starting blueprint: 

• Approach AI from a trustworthy perspective and ensure this starts at the top of the organization with your board.   
• Prepare for regulations and law changes, and enact principles now. Educate yourself about these regulations and they might mean to your company.
• At some point, be prepared to dispose of the data you’re collecting. This is particularly important for retail, where obtaining data from consumers is used to shape personal experiences for them. 

How AI can make supply chains more sustainable 

Though there’s always a cost in investing in new technology, the benefits of AI in the supply chain can translate into long-term gains by improving productivity and reducing waste. It can also supply visibility into nearly every link in the chain — quickly analyzing data sets to provide a deeper understanding of the best place to source raw materials and/or partner with ethical suppliers. 

In fact, high-performing supply chains are 19% on average more likely to have capabilities in place to achieve their sustainability goals, according to Gartner. 

“Through digital transformation … and making it visible and usable through analytics or artificial intelligence in real time to all parties that need it, organizations can make faster and better decisions that lead to improved business and sustainability outcomes, such as the rightsizing and appropriate allocation of inventories and the reductions in fuel, energy, packaging, and other material usage,” says Jordan Speer, research director of product sourcing, fulfillment, and sustainability at IDC. 

Among the ways that artificial intelligence can support a more sustainable supply chain include:   

Fulfillment: Applying artificial intelligence and advanced analytics across a fulfillment strategy can help manufacturers identify the ideal location and method to pick, pack, and ship orders to retailers and consumers. In doing so, they can potentially reduce their emission by ensuring vehicles are traveling the shortest distance that’s necessary.  

Energy efficiency: Artificial intelligence can analyze energy usage in factories to help identify where machinery is being underutilized. Some companies also develop AI-infused digital twins, which are virtual replicas of products and processes, to proactively predict when their equipment may need to be repaired or replaced.  

Demand forecasting and planning: Fewer goods made, fewer goods tossed. Gathering and crunching internal and external forecasts can provide companies with a more accurate estimation of the number of products they should manufacture, helping to reduce the development and disposal of unwanted and unsold items. 

Inventory management: The benefits of this visibility can extend into inventory management. Armed with the knowledge of the ideal location for their products to be stored, shipped, and sold, manufacturers and retailers can optimize their retail strategies to reduce waste. 

Transportation optimization: AI in transportation can be used to identify and optimize delivery routes to reduce the amount of fuel used, decreasing harmful emissions. 

What is an example of AI in logistics?

Unilever is employing satellite imaging, artificial intelligence, and geolocation data to advance its goal of having a deforestation-free supply chain for palm oil, paper and board, tea, soy, and cocoa. The company has partnered with Google Cloud to leverage nearly 40 years of continuous satellite imagery and machine learning to monitor mills, landscapes, and farms. This is expected to enable it to estimate which farms and plantations are supplying the mills. 

Another example of artificial intelligence in logistics includes shipping and logistics company Maersk, which is using artificial intelligence to build a “predictive cargo arrive mode,” according to CNBC. Maersk expects the investment to improve scheduled reliability for its customers. 

“Reliability is a big deal, even post-pandemic, so that they can plan their supply chain, their inventories better, and bring their costs down,” Navneet Kapoor, chief technology and information officer, tells CNBC.

Which companies use AI in supply chain? 

There are many companies using AI across the supply chain, but they are far from the majority. Eighty-seven percent of retail executives in a 2022 Incisiv survey said they hadn’t taken meaningful steps to embrace AI. A scant 6% described themselves as early adopters, while another 8% have dabbled or enjoyed early success with piloted technology. A fair amount (39%) indicated they’re unsure how to apply AI to real-world scenarios.

The examples of artificial intelligence in supply chain management — and otherwise — range the gamut. 

PepsiCo: The food and beverage company leverages artificial intelligence monitoring for predictive asset maintenance, employee safety, and quality in its factories, warehouses, and distribution centers. In doing so, they’re not only protecting their workers but also providing them with visibility into how they can operate more efficiently. 

Unilever: Its laundry detergent plant in Indaiatuba, Brazil, leverages AI, machine learning, and digital twins to predict new processes for laundry powder formulas. It’s not only improved cost efficiency and agility, but it has also eliminated the need for physical trials to quicken product innovation, as well as reduced the company’s environmental footprint. 

Tyson Foods: The food manufacturer has invested in supply chain AI startup Soft Robotics, which aims to reduce labor shortages in food processing environments via robotic picking solutions using AI, 3D vision, and the company’s patented soft grasping technology. 

Proximo Spirits: The spirits company, which is a subsidiary of Becle and includes such brands as Jose Cuervo, Three Olives, and 1800 in its portfolio, is investing in decision intelligence to improve its forecasting and demand planning capabilities and says it’s already improved its forecast accuracy by 30% in the United States. Decision intelligence employs both AI and machine learning to help companies improve their decision-making effectiveness; the spirits company is using it to better predict demand and manage inventory. 

“Most companies need to be very open and very creative when exploring these technologies,” says Luis Gonzalez, Becle global supply chain director. “Because at the end, those that are able to discover the additional benefits that you can get out of it will be the ones that will get the competitive advantage in the future.” 

Procter & Gamble: The CPG is using AI, ML, and edge computing to digitize and integrate data from 100-plus manufacturing sites around the world, leveraging data storage platforms to improve energy use across its paper machines by determining which machines require maintenance and where they can reduce energy use. It’s also using AI to detect changes in tree cover to determine the amount of carbon stored by forests and the associated climate threat if trees are removed. 

Colgate-Palmolive: Mark another CPG turning to decision intelligence. Colgate-Palmolive is using the tech across its Hill's Science Diet and Hill's Prescription Diet fulfillment network. Designed to automate decision-making and forecast demand to precisely designate optimal product allocation, it’s expected to improve stock deployment operations across the company’s fulfillment networks. 

Coca-Cola: Coca-Cola İçecek (CCI), an Istanbul bottler, built a digital twin of its manufacturing plants that used AI and advanced analytics to proactively identify machine failures. 

What is the future of AI in supply chain management? 

Predicting the future of AI can be exciting. If you talk to many manufacturing leaders today, it’s not uncommon for their eyes to light up when they discuss all the ways they envision AI driving efficiencies across their business. 

While rabid enthusiasm should be avoided lest leaders fail to properly account for the host of challenges discussed here, AI can serve as a key component of supply chain progress. 

“Like any industry, the most successful CPG businesses exhibit excellence in supply chain management,’’ says Tyson Foods’ Clark. “Access to near-real-time actionable insights and machine learning-driven recommendations and alerts is crucial to achieving this. Companies that are not investing in this capability will soon find it difficult to compete.”

As such, when exploring the future of AI in the supply chain, it makes sense to look at generative AI, a subset of artificial intelligence. 

Generative AI will support supply chain management in a more sophisticated and reliable capacity as large language models are continuously developed and honed, says Blue Yonder’s Ghosh. In just a few years, we may even see supply chain teams trusting their “AI-powered co-pilots” to take actions against specific problems identified in patterns, such as resolving shipment delays through a specific action. 

“Rather than interrupting their workflows and working side-by-side with their co-pilot on a repetitive sequence of troubleshooting steps, supply chain planners can simply delegate the planning to their co-pilot and grant final approval when the plan is ready for execution,” he says. “This will save supply chain teams countless hours and will be a tremendous boon to productivity and response times.” 

Ghosh believes that generative AI is one of the technologies that will transform the supply chain by adding predictions and intelligence to combat inefficiencies and latency. 

“The more we can develop and implement this technology, the closer we can get to a world with no shortages; a world where sustainability doesn’t come at the expense of convenience; a world where the local price is the best price.”

How AI is changing supply chains

The use of artificial intelligence, while not yet mainstream, is nonetheless effecting change within the supply chain. The consumer goods supply chain continues to grow in complexity, and AI can facilitate a more collaborative and resilient supply chain — something the future of the supply chain depends upon.   

“S&OP has changed significantly over the past four years with COVID, and big shippers are now using AI to understand the likelihood of different outcomes given data like customer demand coupled with macroeconomic viewpoints,” says Keith Moore, CEO of AutoScheduler.  “Instead of just creating single-number demand forecasts, these probabilities identify the range of outcomes that may occur and assign probabilities to each for supply chain teams to war game around.”

What’s more, pairing technologies like machine learning with more traditional optimization techniques can maximize desired outcomes in logistics by modeling constraints and trade-offs, he adds.

Will the supply chain be replaced by AI?

We’re a long way from relinquishing our supply chain duties to AI and machine learning. Even those who see great promise for AI in the supply chain point out that it’s simply another tool, albeit one that can help optimize many  tools.  

Ghosh cites the still-developing quantum computing as a prime example. “We can imagine businesses will use generative AI to help them fully harness quantum’s power,” he says, “but AI could never replace quantum computing. Ultimately, we think AI can help bridge the gap between humans and other advanced technology.”