Intelligent enterprise fueling the supply chain of the future has been saved
Cover image by: Emily Moreano
Black swan events—popularized by statistician and risk analyst Nassim Nicholas Taleb—describe extremely disruptive outlier events that can’t be anticipated scientifically. In recent years, global supply chains have been hit by a series of such black swans: acts of terrorism, armed conflict, political upheaval, port strikes, rise in tariffs, intellectual property risks, and, of course, natural disasters. Most recently, large-scale shutdowns, mobility restrictions, and labor shortages triggered by the pandemic created a serious supply chain crisis of increased demand (recall the panic-buying of toilet paper) and the inability of businesses to meet that demand (port congestion due to lack of labor).
Existing supply chains were not designed to deal with such disruptions in the first place. Since the end of World War II, most supply chain strategies have adopted the credo of “just-in-time” to reduce operating costs and maximize efficiency. This essentially involves using information, planning, and forecasting systems to guide global supply chains and effectively align resources with predicted demand signals. However, after living through recent shocks, many companies have started looking for a new credo.1 The traditional “just-in-time” approach for supply chain design may soon be giving way to “just-in-case” or even “just-in-worst-case.” In other words, with the traditional model proving inadequate to withstand disruptive events, companies now want agile and resilient supply chains.
Reaching there would need organizations to rethink the relationship between supply chain strategy, operations, human beings, and the available technology. Leveraging advanced digital technologies—which can now detect, analyze, predict, and provide prescriptive options in ways that legacy systems cannot—will change how supply chains are planned and executed, and how organizations and roles are structured.
This new vision will enable a future where machines will perform majority of the analysis, freeing up managers or other human workers to make strategic decisions, resulting in more resilient supply chains. We’re approaching the dawn of the “intelligent enterprise”—an organization that integrates an ecosystemwide information layer and cutting-edge artificial intelligence (AI) and machine learning (ML) algorithms to automate and optimize supply chain decisions.
Global supply chains are under intense pressure to rapidly transform so as to function in a complex world of increasing risk and uncertainty. COVID-19 highlighted how unprepared organizations were dealing with operational, structural, cultural, and data-flow constraints of the legacy information systems that global supply chains still rely on.2 Enterprise resource planning (ERP) for these legacy systems was not done for the unpredictable world of today. It was instead done to handle execution elements and connect data on basic transactions, and material- and production-capacity tasks across functional areas such as manufacturing, finance, procurement, and order management. Such ERP systems thus effectively became the enterprise systems of record.
Over time, as supply chain management became more complex, core ERP functions were extended to advanced planning solutions (APS), which supported sales and operations planning (S&OP) processes. Many companies also extended S&OP to perform integrated business planning (and enterprise planning), which now connects operational planning decisions to commercial strategies and financial objectives. In this way, the simplified vision of an ERP collecting and supporting most key business functions has devolved into a complicated web of processes and technologies addressing the complexities introduced by the ever-evolving supply chain requirements.
An agile response to supply chain disruption, by contrast, is based on the premise that operators with the right information in the right format can make faster, smarter decisions. Legacy systems lack this capability: These systems are often loosely coupled, with data flowing interdepartmentally in a cascading manner. Consequently, data is either late, unintelligible, or wrong, leaving supply chain managers scrambling for insights and foresights. Moreover, it diminishes the range of knowledge upon which decisions are made—decisions that can have implications throughout the extended enterprise. Where legacy data ends up being most timely and accurate is at its point of origin, which only reinforces localized, siloed decision-making.
As recent disruptions made supply unreliable and demand unpredictable, even the mightiest of retail giants stumbled.3 And as social-distancing mandates turned more stringent, customers stayed away from physical stores. In response, retailers reduced their field staff and were often understocked due to supply issues. Customer preferences, meanwhile, changed overnight and curbside pickup service emerged as a major imperative.
For a global apparel manufacturer and retailer we work with, these fluctuations in demand, staffing, and inventory ripple throughout the enterprise. What might have seemed like a good planning idea in one functional area may be detrimental to another. This company sells multiple brands of apparel, footwear, and accessories into over 1,000 company-owned stores worldwide. It sources hundreds of millions of units of apparel and accessories annually. Various products and multiple market channels inevitably foster an increasingly complex supply chain, with hundreds of global vendors that operate hundreds of factories worldwide.
In an attempt to decompose its own internal siloes, the retailer tried a “balanced” approach between internal functional areas using a structure comprising “vertical” and “horizontal leads.” Vertical leads focus on specific brands and advocate for their need within the supply chain, while horizonal leads perform functional roles to drive scale and leverage. This helped the company move out of independent silos. However, while balancing the interests of specific functional areas, this approach still falls far short of accounting for the enterprise as a whole.
This dynamic—between vertical and horizontal leads—has profound implications, as organizational design often reflects the impact of legacy-system structures on functional operational alignment (e.g., planning, procurement, manufacturing, finance, etc.).
Further, functional data feeds functional key performance indicators (KPIs), and as a result, product and service delivery decisions are made via competing functions with independent incentives and objectives. This misalignment in performance management leads to functional tension as leaders often chase competing priorities, and time is wasted in overcoming internal constraints (such as influence exerted on internal stakeholders to help fulfill a leader’s KPIs) rather than working collaboratively toward a single, customer-centric objective set at the enterprise level.
It’s now clear that legacy system–driven supply chains are woefully inadequate to tackle future disruptions. A structure designed to the rule will struggle to respond to the exception. Further, it has been sufficiently demonstrated that advanced digital technologies can play a significant role in operational management.4 Consider the ability to capture big data and use AI and ML to quickly identify critical information and offer prescriptive options. Or the use of the Internet of Things (IoT) to remotely manage automated factories. Or digital twins to simulate functional operations, or robots to efficiently perform the most arduous of those functions.
With uncertainty becoming the new certainty, the bifurcation of organizations that will thrive (versus those that won’t) will likely rest largely on their ability to adapt to this new environment. Clearly, businesses must and are already implementing digital technologies to do more of the work that has been traditionally performed by legacy systems and human beings. But what is even more pressing—and perhaps more difficult—is the effort required to create a governance and operational model that is aligned to leverage these new technological capabilities.
Getting there calls for developing digital models that can simulate how any change in supply networks, product design, sales opportunities, or customer mix would impact the entire enterprise. This new model will assume that the organization operates by the use of data models, drawing digitally stored information from the cloud. Then, AI, ML, and other digital algorithms can help identify problems before they occur, and run multiple simulations based on business objectives and priorities to determine potential consequences and trade-offs throughout the enterprise.
This may sound like an enterprise on steroids being manipulated in an imagined laboratory, where virtually, anything is possible. But that is the objective—organizations eventually transforming into intelligent enterprises that adopt advanced digital technologies to create an agile supply chain model that is no longer defined by functional silos (figure 1).
An intelligent enterprise will require the following components—an insights and decision platform, a digital organization, and a digital operating model.
To make intelligent decisions in large, complex supply chains, operators should understand how their actions anywhere will impact the enterprise everywhere. Disruptions in one area could mean delayed or canceled orders in another. Accurate decisions that impact the production, sourcing, or routing of products must be based on data that is timely, accurate, pertinent, and holistic. Many of the business rules and algorithms will likely lead to process automation.
Process automation enables managers to make decisions based on the rapid analysis of potential decision points and how those will impact the entire organization. To do this, the insights and decision platform must include the technology and tools necessary to access both internal and external data sources, captured in real time and organized into a single data model, made available to supply chain operators through a rules-driven, cloud-based layer that spans the enterprise. AI/ML capabilities will subsequently be used to conduct in-the-moment analysis. The findings can then be used not only to respond to supply chain exceptions, but also to guide the identification and automation of manual processes and non–value-add decisions.
To bring the organizational model up to speed with the capabilities of advanced technologies, the organization needs to operate from one integrated data layer, pulling data into one rationalized dataset, accessible to the entire enterprise. AI and other sophisticated analytics solutions, meanwhile, will expedite problem-solving. The result: What would have previously taken days to perform manually within one department will now be executed in a matter of minutes using data from across the organization.
For example, the aforementioned apparel manufacturer accelerated its go-to-market process by utilizing 3D applications to render virtual representations of planned products. Previously, it had used a process that required shipping product prototypes back and forth for approval, which was slow and cumbersome. The new 3D prototypes can be pilot-tested internally and with customers. The resulting digital product development process will get products to market faster, while also speeding product life cycles in markets where consumer preferences change or require regional adjustment.
Liberated from legacy systems, organizations can construct a governance, technology, data, and operational model to support an agile business. The company mission becomes customer-centric, and the business is restructured to provide the resources to nimbly pursue that mission.
Moreover, organizations can be structured to prevent and address service failures. Execution-focused roles in business functions would deliver customer-centric solutions using systems and analytics from cross-functional pods. Blending functional roles, such as planning and execution, can improve collaboration and responsiveness when facing near-term market fluctuations and disruptions. For example, positions such as a Master Planner (historically focused on monthly/weekly supply plan in an APS system), an Inventory Analyst (focused on inventory management), and a Replenisher (focused on fulfilling near-term demand in an ERP or execution system) will be combined into one role. The consolidated function will utilize the same data, business rules, and policies to respond to demand changes and iteratively replan with the latest input from the market. Reducing the number of functional roles and systems involved increases the organization’s agility and decision-making speed.
The digital operating model will become operational through a series of codified business rules that will support the new customer-centric, rapid-response approach. The goal of the operating model is to help identify and address the root cause of problems. In effect, this model assumes that supply chain disruption is an eventuality, rather than an exception.
The design of the operating model will assign decision-making authority based on information and analytics. Where formerly the governance model supported legacy system–dependent business silos, management can now replace those with policies, procedures, and incentives based on the integrated business planning process (i.e., the enterprise business planning process).
The technology layer provides a common, integrated data repository (single version of the truth) that will be accessed by roles across the enterprise to help ensure enterprisewide impact. In this model, performance measurement and incentives will be associated through a common, holistic set of metrics. For example, the aforementioned global apparel manufacturer is transitioning its assortment planning, allocation, and retail planning to a digital platform. When completed, this platform will help provide a single view of end-to-end supply chain operations across its global network.
As the company transforms, so does the work. New digital technologies will help automate mundane, repetitive, transactional, and non–value-added work (e.g., manually manipulating spreadsheets to connect datasets, or extracting data from siloed legacy systems). New roles will be focused on capability (e.g., data science economists responsible for weighing strategic and financial trade-offs and managing the business policies and logic in the digital layer) rather than functional expertise. The work will now be more strategic since it is focused on data analysis, running scenario modeling, and debating the merits of various strategic choices amplified by new digital tools.
Responding deliberately to disruptions requires interconnectedness of all enterprisewide variables. These include commercial strategies, production capabilities, financial impacts, inventory and supplier visibility, and order management capabilities (as the interface to the customer). Unfortunately, no single system does this today. Meanwhile, the speed to evaluate and make decisions is not sufficient to respond to ever-changing customer needs. The existing operational model thus does not allow for necessary transformation.
But like any transformation, this one too starts by acknowledging that the status quo is no longer tenable. Then, objectives that will get the organization to cultivate a new customer-centric, extended enterprise–aware culture can be outlined. What kind of organization do you want to be? What is your identity? Who should be driving the final decisions?
To answer these questions, companies must figure out how human talent and digital technologies can work together (augmentation). Machines will do more of the work in the future, which can yield returns in the short run (reduction in head count and fewer legacy systems to manage) and provide significant benefits into the future (improved decision-making driving enterprise priorities, improved agility/resilience).
Once organizations have set their objectives, the next step will be driving toward a single system to collect data from all functions across the enterprise and consolidating it into one data source (the single version of truth). This will require a system layer that connects internal and external data sources, likely stored in the cloud.
Once the single data layer has been established, organizations will be in position to use advanced digital technologies to optimize decision-making. AI/ML capabilities can automate low-impact or time-consuming decision-making by learning from the impact of previous decisions, while simultaneously refining the algorithms. Staffing can be reorganized to leverage analytics that will focus the enterprise on customer-centricity rather than functional silos.
Clearly, organizational change on this scale is not going to be easy. Nor will it happen overnight. There’s no single right answer for how to reach the intelligent-enterprise end state. Many organizations have embarked on this transformative journey, but few, if any, have mastered it. In fact, most leaders are currently focused on laying the foundations.
For example, the aforementioned global retailer has already launched its intelligent-enterprise efforts by developing a new vision for better decision-making in an uncertain environment. The first step in that vision is modeling the decisions made across the entire enterprise—from the factory floor to the store. That goal is close to realization, as the retailer has made data connections across the many functions of the enterprise, which will provide a holistic visibility into the supply chain. This resulting data will be used to build decision models, which will ultimately be automated.
The next step is evolving the company culture in alignment with its technological and operational capabilities. Governance models will need to establish clearly defined rules and roles to facilitate in-time decision-making. Skill sets will need to be upgraded, and operations managers will need to perform data-driven analysis for customer-centric decisions (with enterprisewide impact) from a single version of the truth while building and analyzing simulated decision scenarios.
In closing, it is increasingly clear that for supply chains globally more black swan events and disruption lay ahead.5 This accentuates the need for organizations the world over to transform into intelligent enterprises. Many companies with whom we work in the retail and consumer goods manufacturing spaces have already begun laying foundational elements. The race is on to improve operations, deliver better services, and offer improved value with agile and resilient supply chains.
Adam Mussomeli, Paul Delesalle, and Jim Kilpatrick, The new supply chain equilibrium, Deloitte Insights, April 1, 2022.View in Article
Lis Evenstad, “Legacy data and IT issues ‘laid bare’ during Covid-19, says NAO,” Computer Weekly, May 19, 2021; Wu He, Zuopeng (Justin) Zhang, and Wenzhuo Li, “Information technology solutions, challenges, and suggestions for tackling the COVID-19 pandemic,” International Journal of Information Management 57 (2021).View in Article
Deloitte, “The retail evolution’s great acceleration: How to maneuver in the pandemic-driven recession,” accessed November 11, 2022.View in Article
Nitin Mittal, Dave Kuder, and Samir Hans, AI-fueled organizations: Reaching AI’s full potential in the enterprise, Deloitte Insights, January 16, 2019.View in Article
Willy C. Shih, “Global supply chains in a post-pandemic world,” Harvard Business Review, September–October 2020.View in Article
Cover image by: Emily Moreano