In the past decade, digitalization and artificial intelligence (AI) have transformed the competitive landscape in many industries—and this revolution will only continue. One of the most visible examples is the automotive industry, where autonomous product functionality is quickly evolving from cruise control to self-driving vehicles. Supported by AI, autonomous systems in cars are able to gather a wealth of data, process it, and respond much faster and more reliably than humans with limited cognition. These advanced autonomous systems can make informed decisions that protect passenger safety, conserve fuel, reduce emissions, manage maintenance, and maintain comfortable conditions inside the vehicle.
Because automakers were among the first to recognize and embrace the benefits of autonomy for their product designs, it's no surprise that the world's car manufacturers are also at the forefront of supply chain automation. Companies like Mercedes-Benz USA and Renault are pioneering the concept of the autonomous supply chain, in which any deviations from planned performance are sensed, analyzed, and corrected in real time, automatically and without any human involvement. Organizations can now increasingly harness vast amounts of data from both digital and physical assets into their supply chain business processes and leverage AI and machine learning (ML) to make faster and better decisions for superior customer experiences and a competitive advantage. By applying AI and ML to this data, companies can gain "predictive visibility," or the ability to sense a disruption before it happens. Once organizations have sensed this potential disruption, they can then use prescriptive analytics to create appropriate risk mitigation plans. Organizations must recognize they need to radically transform their operating models to achieve greater visibility, sense and predict disruptions more accurately, and respond with speed and agility—while also protecting profit margins.1
![[Figure1] Three crucial supply chain evolutions in the automotive industry](/media-library/20191214auto-ex1.jpg?id=53611656)
![[Figure 2] Roadmap to the autonomous supply chain](/media-library/20191214auto-ex2.jpg?id=53611657)
Automakers are natural champions for supply chain autonomy. They face continuing pressure to launch new, complex product designs (such as electric vehicles that meet strict global emissions deadlines) faster than ever before, and they operate in geographically sprawling, complex supply chains. Only artificial intelligence has the ability to consider conditions along the end-to-end automotive supply chain, perform real-time analysis, and automatically make decisions that can keep this entire network on track.
Sweeping changes
In addition to their ongoing need to work efficiently and manage extreme supply chain complexity, automakers have been impacted by three recent trends that make autonomous operation even more critical. These trends add up to a need for new levels of predictive visibility, responsiveness, and customization that can only be achieved via AI and machine learning (see Figure 1). These trends include:
- A shift from a node to a network mindset for risk mitigation. The auto industry is a complex ecosystem of dealers, original equipment manufacturers, tier-1 and tier-2 suppliers, third-party logistics providers, and electronic manufacturing services. Any disruption in one node of the supply chain—for example, a supplier parts shortage—can halt production and interfere with delivery deadlines. This actually happened to Ford in May 2018 when it was forced to temporarily halt production of its popular F-150 pickup truck in Dearborn, Michigan, and Kansas City, Missouri, because of a supplier parts shortage. It is important, therefore, to be able to identify risks and deviations not just at the node level but across the entire network. Artificial intelligence makes this now possible. It enables this new level of predictive visibility, allowing issues to be identified and communicated across the entire end-to-end supply chain. Armed with AI, automakers can act rapidly to mitigate the negative impacts of any disruption, anywhere in the value chain.
- A transformation of planning horizons from periodic to real time. A decade ago, functionally siloed, batch-level planning was the norm, as automotive consumers' service and delivery expectations were lower. But today the "Amazon effect" has trained consumers in every industry, including automotive, to expect high levels of responsiveness and customer service—whether they are ordering a new vehicle or spare parts. Monthly, weekly, or even daily planning is no longer enough. Now automakers need real-time, "always on" planning capabilities that eliminate the latency between sensing consumer demand and responding to it. Removing a day of data latency can mean an increase in top-line revenue of several million dollars as well as bottom-line cost savings because of working capital reduction—and only AI and autonomy can support this level of real-time responsiveness.
- An evolution from uniform products and processes to personalization. Auto manufacturing has traditionally followed a "push" make-to-stock (MTS) model, based on long-term demand forecasts for uniform product designs. Today's consumers, however, crave product customization. As new smart features, electric powertrains, and hybrid models are introduced, and consumers have more options to choose from, this trend will only increase. Across the end-to-end supply chain, all stakeholders must now be able to participate in a "pull-based" assemble-to-order (ATO) or configure-to-order (CTO) environment that customizes products and processes to meet specific consumer needs. Only those OEMs that can accurately identify consumer needs and intelligently configure (and reconfigure) the supply chain will be profitable—mandating the need for process automation and autonomous decision-making via AI.
Bottom line, there has been an exponential increase in automotive supply chain complexity because of these three trends. And automakers are not alone. In a recent study sponsored by JDA Software and KPMG, manufacturers across industries named these three key drivers of investment in their supply chains:2
- Real-time product visibility (named by 50% of respondents),
- Greater speed and agility (40%), and
- Increased automation and lower total cost of ownership (33%).
In the past, organizations have been able to survive extreme complexity and volatility by throwing inventory, capacity, and planners at any performance deviation. However, with shrinking margins, cutthroat competition, and more demanding consumers, this strategy is no longer sustainable.
The good news? Artificial intelligence and autonomy can help companies respond to this increasing complexity, and achieving autonomy is increasingly within the grasp of the world's automakers and other manufacturers. Not only can AI gather and interpret real-time data from across the supply chain via a shared technology platform, but it can also incorporate third-party, nonstructured data such as weather forecasts, social media trends, and news events—all of which can significantly affect both supply and demand. By leveraging prescriptive analytics, AI can then transform these insights into strategic actions that maximize speed, agility, and customer service—as well as profit margins.
The four levels of supply chain autonomy
So how can the world's auto manufacturers and their tier-1 and tier-2 suppliers make this shift from their traditional business models and processes to the autonomous supply chain? Just like the development of autonomous vehicles, it is a step-by-step journey. When it comes to autonomous cars, the five steps toward full autonomy include:3
- Level 1: Feet off, which involves the use of cruise control to manage acceleration.
- Level 2: Hands off, which means the car automatically maintains a safe following distance or stays within lanes.
- Level 3: Eyes off, which allows the driver to read or do other activities—but still intervene in an emergency.
- Levels 4 and 5: Mind off, which translates to the fourth step, high automation, and finally, the fifth step, which is fully autonomous driving, with humans as passengers
Similarly, there are four steps in progressing toward the fully autonomous supply chain (see Figure 2).
Level 1: Increased end-to-end visibility. The autonomous supply chain starts with achieving real-time visibility into "where stuff is" across the end-to-end supply chain. This includes inventory, spare parts, and other assets at supplier locations, factories, warehouses, distribution centers, and dealerships. This increased level of visibility is achieved by mining data from multiple enterprise resource planning (ERP) systems, warehouse management systems (WMS), transportation management systems (TMS), and advanced planning and scheduling (APS) systems. Having suppliers connected on a common technology platform and utilizing common software solutions are essential in making this level of visibility possible.
Level 2: The use of predictive analytics. Once companies have visibility into their assets, they can move on to sensing and predicting any performance deviations before they happen. Only artificial intelligence has the capability to analyze data from across the supply chain; incorporate it with third-party, unstructured data; and apply sophisticated algorithms and analytics to sense changes before they impact performance.
Level 3: Automating corrective actions using prescriptive analytics. Once disruptions are predicted, AI-supported digital "control tower" technologies have the power to recalibrate plans in real time, using what-if scenarios to mitigate risk and maximize financial outcomes. Backed by AI and machine learning, these digital control tower solutions can make trade-offs, balance outcomes, and respond with extreme speed and agility, without human intervention. For example, in the event of a missed spare part delivery, digital control towers can weigh such options as expediting a replacement from the original location or shipping from an alternate location. They can evaluate the speed of different carriers and routes to minimize any negative impact caused by the disruption.
Level 4: Achieving true self-learning. In this advanced stage, companies use machine learning to identify more subtle, complex patterns in the collected data, based on historical disruptions and their eventual outcomes. Artificial intelligence and machine learning engines work together to establish correlations and patterns, so they can recognize an event far in advance, based on lessons learned from the past. This "self-learning" supply chain actually automates the process of continuous improvement, becoming smarter and producing better outcomes all the time.
While the vision of the truly autonomous, self-learning supply chain may seem ambitious, a number of automotive companies have already embraced this concept and are working to bring it to fruition.
For example, in partnership with JDA, Renault is working to automate its sales and operations planning (S&OP) process through what it calls its R3 project: "Right Car, Right Time, Right Place." Florian Huettl, global vice president at Renault, described the program in a recent interview: "My forecasting managers spend too much time doing jobs that could be automated. And they don't have enough time to look at the actual markets, to get a deeper understanding of the market itself. We started using automatic forecasts [and] AI. We're doing it now on markets and sales potential. It's still early days, but we are using the data, and it's part of our forecasting process. We're covering short-term horizons, so we have started to have machine learning, predictive forecasting, [looking at] the next six months. We are looking at some kind of autopilot approach where, in fact, we can delegate tasks of forecasting to algorithms, to our machine learning suite, and the job for AI will be to do the forecasting within limits that the forecasting managers will set. We're trying to teach our algorithms to 'raise a hand' when they see that we're running into trouble, and it goes beyond their capacity."4
Mercedes-Benz USA's journey
Mercedes-Benz USA (MBUSA) is also embarking on a strategic journey toward an autonomous global supply chain. MBUSA is a distributor for passenger cars in the United States and the second-largest car sales organization of Daimler AG. Mercedes-Benz USA's Parts Logistics organization handles forecasting, fulfillment, and transportation for all aftersales parts for its vans and passenger cars.
Its uniquely complex supply chain model makes the Parts Logistics organization well-positioned to capitalize on the benefits of using artificial intelligence and automation to respond to supply chain disruptions. MBUSA sources approximately 90% of its stock-keeping units (SKUs) from Daimler AG's Global Logistics Center (GLC) in Germany. The parts travel by sea freight to eight parts distribution centers (PDCs) across North America before being transported to dealerships. Overall, MBUSA plans for several hundred thousand active SKUs for over 550 dealers.
Historically, MBUSA had centrally planned parts for only the United States at its headquarters in Atlanta, Georgia. In 2016, Canada parts planning was shifted to MBUSA, and Mexico followed shortly after, resulting in all of North America being centrally planned in the United States. In the past, MBUSA has struggled with planning due to long lead times, poor transportation visibility, and a parts strategy that is more reactive than proactive. These issues have resulted in modest service levels, high levels of safety stock, and high costs for expedited transportation such as air freight. With rising customer expectations and a network of growing complexity, MBUSA needed a new solution that would provide the predictive visibility, speed, and automation it needs to succeed in today's automotive marketplace.
As a result, MBUSA has embarked on a transformation journey and is partnering with JDA to create a more autonomous supply chain in three key ways:
1. Planning. To set up is supply chain for success, MBUSA believes that it needs anintegrated business planning process that incorporates a concern for data integrity, strategic initiatives, financial guidelines, and market intelligence. An aligned S&OP process and data management program builds a strong foundation for reaching an organization's fulfillment goals. Before it can implement a digital solution for handling supply chain disruptions, a company needs to create a strong planning foundation.
2. Sensing potential disruptions. Without visibility and synchronous information sharing, the business cannot proactively respond to disruptions in the supply chain. For example, poor weather or port congestion may delay a shipment of hot items from a supplier to a parts distribution warehouse, but the estimated time of arrival (ETA) in the ERP and supply chain planning systems will not be updated to reflect this new information. This data latency manifests itself in stockouts, referral costs, and service-level hits. JDA's digital control tower can provide real-time shipment visibility and predict possible delivery disruptions or delays. By implementing this solution, MBUSA will gain the ability to detect changes in ETAs, fill rates, and inventory levels in real time. Exceptions will create automatic alerts to users, notifying them of shipping delays or inventory-level changes that may require attention.
3. Responding to disruptions. MBUSA wanted a digital supply chain solution that would help it sense and respond to risks and opportunities in its supply chain. JDA's digital control tower, which has planner exception management and prescriptive corrective actions, will help with this. When a shipment is delayed, machine learning is leveraged to create predicted time of arrivals (PTAs) for shipments. These PTAs trigger automated alerts to planners. The planner can then respond to these delays by utilizing an alternate sourcing lane to restore inventory levels. Similarly, if the control tower projects a stock shortage, it can trigger a more aggressive fulfillment strategy for the affected parts.
The journey toward an autonomous supply chain starts with having a business-use case. The MBUSA Part Logistics team focused on tying the business' current challenges to the benefits that an autonomous supply chain could bring. Once it identified the current-state challenges, MBUSA identified who the primary users of the digital control tower would be.
Representatives from these groups—such as the planning and replenishment team, the PDCs, transportation team, and the dealer network—were then invited to participate in workshops to define the key enablers for their teams. This process garnered buy-in and resulted in a list of "must have" capabilities such as track-and-trace with predictive ETAs, single application source alerts, improved data integrity and accessibility, recommendation of corrective actions, and improved container prioritization.
MBUSA is still in the early stages of its autonomous supply chain journey, but the company expects the benefits to include higher service levels at a lower inventory cost, reduced operational labor hours, and the ability to appropriately manage dealer parts ETA expectations.
Just the start
While a truly autonomous, self-learning supply chain seems futuristic, best-in-class companies, like MBUSA, are making significant progress toward this goal. As demonstrated by the Mercedes-Benz example, enabling the autonomous supply chain is not a sudden, all-or-nothing event. Instead it is a journey, where the company is leveraging and adopting advanced technology in the areas where it makes the most strategic sense to solve real-life business use cases.
Technology has come a long way, and advancements in AI and ML make it possible for companies to analyze vast amounts of data to predict events, spot trends and anomalies, analyze any correlations between patterns and variables, understand the ramifications of different response options, and provide recommendations. Now with the ability to analyze outcomes (including consumer insights and past behavior) and adjust supply parameters, companies can make faster and better decisions and provide better visibility, speed, and automation to address the growing complexity. And they can do so profitably. By leveraging the power of AI, companies can now take advantage of the digital revolution to create a fully connected, automated supply chain that gathers information, performs sophisticated analysis, and takes strategically correct action with no human intervention.
Whatever a company's current stage in its supply chain evolution, it is essential to remember that achieving full autonomy is an ongoing process, beginning with single technology implementations in critical areas of the business. But the end benefits—including increased supply chain visibility, reduced expenses, higher performance, increased efficiency, revenue growth, improved margins, and higher shareholder value—make the journey to full autonomy worthwhile.
Notes:
1. Salim Shaikh, "Cognitive Integrated Business Planning," Technology Optimization and Change Management for Successful Digital Supply Chains, JDA, 2019, https://www.igi-global.com/chapter/cognitive-integrated-business-planning/223330.
2. "2018 JDA & KPMG Digital Supply Chain in Retail & Manufacturing: A State of the Industry Benchmark conducted by Incisiv," https://jda.com/knowledge-center/collateral/kpmg-jda-digital-supply-chain-in-retail-mfg-executive-summary.
3. Isabel Harner, "The 5 Autonomous Driving Levels Explained," IoTforall.com, Oct. 23, 2017, https://www.iotforall.com/5-autonomous-driving-levels-explained/.
4. Victoria Johns, "Q&A: Florian Huettl," Automotive Logistics (Jan. 7, 2019), https://www.automotivelogistics.media/qanda-florian-huettl/22164.article.
