No one would deny that managing a global supply chain is an increasingly difficult task. Today’s supply chain managers have to contend with significant disruptions—such as those caused by the COVID-19 pandemic, trade tensions, and natural disasters—as well as growing complexity from forces such as omnichannel retailing and increased customization. Many experts believe that to effectively manage this difficult terrain, companies have no choice but to harness the potential of digital technologies such as data analytics, artificial intelligence (AI), and robotic process automation (RPA).
At Auburn University, we have been studying how companies engage in this process of supply chain digital transformation. Toward this end, we conducted a focus group with 15 industry partners and interviewed an additional six individuals who are actively involved in supply chain digital transformation efforts. Our exploration focused on how these organizations deploy and utilize technology to achieve their transformation goals and how they are integrating technology and human resources to address supply chain challenges. (For more information on our methodology, see the sidebar “About this study” below.)
These experts spoke with us about the lessons learned from their digital transformation efforts while exploring what we see as the final phase of the journey: leveraging digital technologies to change the value proposition for the organization and redesigning processes so that the responsibility for execution lies primarily on technology instead of human workers.
We believe that to achieve this final stage, companies will need to embrace what are known as intelligent workflows or the purposeful planning of interactions between humans and technology. Intelligent workflows are a blending of technology and human workers that offer a comprehensive approach to orchestrating automation, AI, human workers, and system integration across entire business processes. They go beyond the realm of human-led AI and simple automation. They place technology at the center of end-to-end process execution, allowing humans to focus on providing high-value subject matter expertise.
What is digital transformation?
Digital transformation is a multiple-stage process by which organizations encapsulate, assess, and then shape their use of data and digital technologies to create additional value for themselves, their partners, and their customers. The process comprises three arguably sequential steps: digitization, digitalization, and digital transformation.1
These steps are often confused with each other, but we define them in the following manner. Digitization, sometimes called digital encapsulation, is converting existing data and documents into a digital format to accurately represent the physical world. In this stage, data is not altered or analyzed, it is merely encoded. Digitalization, predicated upon digitization, is altering processes, organizational structures, or decision-making architectures to leverage improved data capture, analysis, and information dissemination. Finally, digital transformation fundamentally changes the process to fully leverage these new digital technologies within and across firms. Digital transformation ultimately affects how the organization creates value within its supply network. The table in Figure 1 provides more detail about the different phases of digital transformation and the technologies involved.
The majority of the companies involved in our study are in the midst of a digital transformation effort but have not yet entered the final stage. Most had already taken the first step of digitizing their data, seeing it as essential to company survival. As one participant said: “Digitization of the supply chain is a requirement for being able to be an omnichannel retailer in the future. You’ve got to know what you have. You’ve got to know where it is with [a] high degree of accuracy. Or you’re dead.”2
Many of our participating companies had also moved beyond digitizing their data and into the digitalization phase. At this stage, they are using digital technologies to augment business processes but have not fundamentally transformed them. They are providing their human managers with tools—such as dashboards, inventory trackers, alerting systems, and even RPA and bots—to improve the efficiency and effectiveness of their processes. The human managers, however, are still central to the execution of the process. For example, alerts may bring a situation to a manager’s attention when a predefined digital metric is tripped, but it is still up to the manager to act. Similarly, dashboards may be collecting data from multiple sources in one place, but managers still have to interpret and act upon that data.
Only a couple of companies in our study had entered the third phase. One company was actively using intelligent workflows to orchestrate the execution of supply chain processes, and another was building the processes and infrastructure needed to do so.
Limits of digitalization
While many of the companies involved in our study were in the digitalization phase, most of the experts we talked to were already well aware of the limitations of digitalization tools. For example, our experts quickly pointed out that the information presented through sprawling dashboards can be overwhelming for decision makers, who struggled to find the correct information at the point of need. “I was never short of information or dashboards. It was like walking into the Louvre. There's artwork everywhere, but at the end of the day, you walk out like, ‘Wow. Awesome.’ But not sure what the hell to do about it.”
Human managers are still needed to evaluate the significance of the data presented through the dashboards. “We’re already digitized; we keep too much data. We have so many dashboards all over the place that no one looks at them. So, what we’re trying to do is not just be digitized but cognitive in our approach.”
While dashboards and control towers help consolidate information, they often lack flexibility and are only adaptable through human intervention. For example, the human manager/expert asks questions and then uses dashboards and control towers to evaluate additional data before deciding on a solution. The efficiency of this process can be improved through AI, usually in the form of machine learning (ML), to evaluate and combine extensive and complex data to generate predictions based on numerous decision alternatives. But even then, a great deal of human involvement is needed. There are still some number of cases for which the model’s predictions are not correct. Human subject matter experts (SMEs) are still required to focus human attention on cases that would otherwise be mishandled and to generate new training data that can be used to retrain the AI model to improve performance.
Additionally, humans are needed to detect when the conditions under which the model was trained have changed. For example, a machine learning model that predicts transit times trained under pre-COVID conditions would likely perform poorly under disrupted COVID conditions.
Similar limitations exists when it comes to using bots, digital workers, or decision engines to automate traditional human work. This type of RPA (sometimes called intelligent process automation) seeks to automate discrete, repetitive tasks, such as reading data elements from specific, fixed cells within a spreadsheet. Although digital workers are very efficient in carrying out discrete tasks, their ability to complete tasks can vary significantly.
At this point in time, our experts say that automated technology acting independently from human input could only work in specific scenarios. Most real-world processes exist within an extensive, complex context that also involves human users/interactions and integration with other systems. As one participant said, “The prescriptive part is tough because it requires an intimate knowledge of the business you're trying to affect.” A human manager/expert is needed to use their intuitive or “tacit” knowledge in combination with the AI.
Indeed, in most cases, even an AI-enabled control tower still requires a human to orchestrate a business process through its multiple steps. Even in a digitalization environment where existing processes are optimized with technology, there are still steps that require varying degrees of human review/action or the ability to pull data from or push updates to other systems (such as an enterprise resource planning systems). Technology may be able to act independently within the larger process, but a human will still be needed to provide coordination across multiple steps in the process.
However, if this orchestration piece could be automated, even greater efficiencies would be gained. “We have a limited amount of people, and we aren’t going to get more people. That’s the reality. So how do we make the people we have more effective to solve the things we need to solve?”
Digital transformation into intelligent workflows
As described earlier, intelligent workflows refer to the orchestration of process automation, AI, and human experts across an end-to-end business process.
An intelligent workflow implementation plan would provide parameters for the integration of all of the interactions that need to occur among digital workers, human workers, AI, and other IT systems in order to complete an end-to-end business process. Technology would now be handling all the administrative details of ferrying work through the tasks that comprise the business process. For example, RPA (or digital workers) would handle discrete, repetitive, well-defined task work. AI would handle cognitive tasks, such as decision-making and natural language interaction or content capture (for example, extracting information from scanned documents). Finally, human experts would support the overall intelligent workflow through ongoing quality assurance, handling cases that automation/AI cannot manage and investigating/resolving issues where automation/AI is unsuccessful. Human experts would also provide feedback to improve the process automation and AI for continual improvement of the intelligent workflow.
In this scenario, the human worker becomes a supporting actor in the workflow. Although their skills are crucial to the ongoing success and improvement of the workflow, they are not directly responsible for working through the end-to-end business process. That responsibility is now assumed by the workflow orchestration service, the intelligent workflow.
The study participants that are already considering intelligent workflows describe some version of the vision outlined above. Their first steps toward that goal may be similar to the ones that the oil and gas company Shell is taking, as described in a recent Harvard Business Review article.3 Shell has begun reengineering its supply chain, manufacturing, and maintenance processes so that they are enabled by AI. For example, the company is automating its inspection processes, using robots and drones to monitor Shell’s energy and chemical plants, pipelines, offshore facilities, and wind and solar farms. According to the article, “Some Shell facilities are so large that it would previously have taken years to inspect everything manually—now drones and robots are being introduced to automate these processes and help shorten the cycle time.” Human inspectors and technicians play more of a support role, spending their time prioritizing projects, performing more advanced verification, annotating images to improve inspection algorithms, and managing the training processes for ML models.
As they redesign their processes, some of our study participants are also exploring how worker skills will be affected by this transition. Because administrative tasks and workflow management are increasingly automated, the human skills required will focus more on subject matter domain knowledge and process analysis/design.
Some participants reported that workers have said that their contributions feel more significant when their expertise is more effectively utilized. Employees engaged with intelligent workflows may feel their role becomes increasingly strategic and innovative as a result.
Long road ahead
As they described the digital transformation journey, our study participants were clear that the process was lengthy—often lasting multiple years—and involved all levels of the organization. Our experts told us that it is important to take a deliberate approach to transformation that recognizes the importance of employee buy-in and proper encapsulation of data.
As we analyzed the information we gathered from the study participants, it became evident that digitalization improved the speed and quality of decision-making not just by increasing visibility and data sharing across the supply network, but also by improving the decision-making processes themselves. It also was clear that even as organizations used more AI, there is still a key role for human workers. Companies still need to draw on humans’ tacit knowledge to assess the recommendations made by AI and to provide feedback to the process automation and AI portions of the intelligent workflow.
Where do we go from here? Further discussion is needed to investigate how the relationships among partners within a supply network influence the application of technology/information to achieve transparency, security, and responsiveness. Additionally, there is still more to be learned about how intelligent workflows can orchestrate automation, AI/ML (including emerging generative AI technologies), and human interactions across end-to-end processes.
Authors’ Note: We are incredibly appreciative of the insights from our diverse participants. We also want to thank Auburn University’s Center for Supply Chain Innovation and Auburn’s RFID Laboratory for their insights.
Notes:
1. P. C. Verhoef, T. Broekhuizen, Y. Bart, A. Bhattacharya, J.Q. Dong, N. Fabian, and M. Haenlein, “Digital transformation: A multidisciplinary reflection and research agenda,” Journal of Business Research, 122 (2021): 889-901.
2. Quotes throughout this article are from our industry expert interviews and focus group participants.
3. T.H. Davenport, M. Holweg, and D. Jeavons, “How AI Is Helping Companies Redesign Processes,” Harvard Business Review, (March 2, 2023): https://hbr.org/2023/03/how-ai-is-helping-companies-redesign-processes
About this study
Information was collected through a focus group and semi-structured interviews with various leaders in organizations that have experienced unique digital transition initiatives. Many of these leaders spent the years before our project navigating their respective digital transformations and were anxious to share lessons learned.
Expert Panel: A focus group was conducted with 15 industry partners that lasted over an hour. Our questions focused on the most pertinent issues of digital transformation efforts and helped to identify specific areas to explore further in targeted interviews.
Interviews: Interviews were also conducted with six industry participants involved with supply chain digital transformation efforts. These interviews allowed us to dig deeper into the specific and unique thought processes involved in digital transformations. Because this research drew from supply chain experts from a broad array of logistics service providers, retailers, distributors, and manufacturers, we were able to develop an informed perspective on digital transformation practices. Interacting with industry experts led to insights into digital transformation norms and practices.
