Maria Jesus Saenz

Maria Jesus Saenz has been working with the Institute since 2004, though she didn’t arrive in Cambridge until 2018. Prior to assuming her current roles as Executive Director of the MIT Supply Chain Management Master’s Program and Director of the Digital Supply Chain Transformation Lab at MIT, she was the Director of the MIT-Zaragoza International Logistics Center established jointly between the Institute and the Aragonian government in Spain in 2003.

Using quantitative research methodologies to show how data-driven ecosystems create value, Dr. Saenz is particularly interested in the interactions between technology, operations, and strategy from an inter-organizational point of view. She has led various international research projects for the European Commission, as well as for companies on supply chain management innovation such as Coca-Cola Femsa, P&G, Carrefour, Clariant, Dell, and DHL. Current research at her lab draws from her business past and revolves around inter-organizational collaboration while implementing digital technologies with an end-to-end approach in their supply chains.

At MIT, we see academia and industry as inextricably linked.

Today, Dr. Saenz examines supply chains, specifically, the interaction between strategy, operations, and technology. At the MIT Center for Transportation and Logistics, her collaborations with companies from different industries enrich her academic viewpoint while allowing her work to impact industry. “Here at MIT, we see academia and industry as inextricably linked. The companies we work with need academia to innovate,” she says. “At the same time, we learn from real-world business challenges, especially in terms of their inter-organizational approach to collaboration, sharing resources, technology and data to create value driven supply chains.”

According to Dr. Saenz, the supply chain department is well positioned to play an integral role in digital transformation. For starters, supply chain professionals are accustomed to working with data and measuring key performance indicators (KPIs), as a kind of language to communicate with other functions inside and outside the company. Furthermore, as a system, the supply chain integrates so many aspects of business operations, interfacing with commercial departments, purchasing and procurement departments, as well as manufacturing and transportation.

They provide a holistic vision of the entire value chain—from suppliers to customers’ homes. And an integrated understanding is key for innovation through digital transformation because the goal is to structure end-to end data threads that connect all the moving parts. This comprehensive view of the value chain also speaks to the supply chain’s ability to translate different languages within an organization. For example, the supply chain is fluent in translating inventories from financial terms into commercial terms and sales into operations terms.

Recently, Dr, Saenz has been examining how to best combine human awareness and artificial intelligence (AI) to create human-machine teams optimized for successful AI implementation in the supply chain domain. “The first assumption is that both parties, AI and humans, are going to learn through the process—this is essential to accept in order to allow the learning to happen,” Saenz explains.  

By design, an AI-driven system will change as it learns from new data, from the algorithms implemented by humans to trigger learning and improvement in certain performance parameters of the operations. An AI system will inevitably learn from the execution of the previous processes, from its environment, from the information it is gathering in real time, and from human behavior and insights. Meanwhile, humans will learn thanks to AI’s strong data management capabilities. The goal is to facilitate this mutual learning experience, and the aim of Saenz’s research is to prepare the systems to facilitate this process.

But the implementation of AI systems is complicated—the first assumption is that it is going to evolve; it is not static. Which is why Dr. Saenz and her research team, developed a framework to foster successful human-AI teaming. On paper this looks like four quadrants, with one axis representing the process where the AI system will be implemented, and another to illustrate the level of risk of the decisions made by the AI system. Each quadrant illustrates one configuration of how humans and machines can work in tandem, deploying different features or capabilities in terms of human-AI teaming.

In terms of the AI-driven process, there are two extremes: highly structured or unstructured. A highly structured process suggests greater understanding of how to structure the AI and algorithm, which typically means less intervention from the human. At the other extreme, is an unstructured process where the variables and its interactions are not clear for implementing AI. This implies more interaction with the human and greater opportunities for mutual learning.

Referencing her framework, Dr. Saenz notes that when the pandemic hit, uncertainty was extremely high. As a result, risks around decisions were perceived as high. Companies were lost when it came to forecasting product sales because they lacked historical data related to past pandemics, which meant that AI forecasting systems lacked sufficient anchor points from which to learn and therefore predict things like openings and closings of schools or full lockdowns around the globe.

However, businesses that engaged expert forecasters and demand planners to make decisions and compare insights during lockdowns were able to provide anchor points that AI could learn from in order to capture new patterns related to COVID. This helped in two ways. First these businesses were able to improve the forecasting accuracy for certain products by more than thirty percent. Second, the AI system acquired more knowledge for contexts of challenging predictability. The key takeaway is that both actors, human and AI, complemented each other to improve outcomes.

Start the journey of digitalization with the value proposition, understanding what you need to do and why.

Following this track of human-AI complementarity, through her research, Dr. Saenz noticed that while most companies recognized the need for digitalization, many weren’t prepared to tackle the complexity of the transformation. Companies analyzed by the MIT Digital Supply Chain Lab that addressed digitalization from a purely technological standpoint suffered, not knowing why or how to expand the benefits of the technology end to end, and, most importantly, what their future operations would look like. Meanwhile, those that took on digitalization holistically, from a strategic standpoint that considered the vision of their desired supply chains and approached the transformation through organizational and cultural change, improved exponentially and demonstrated a sizable competitive advantage over their peers who were not prepared for the complications inherent in technological evolution.

“My recommendation is to start the journey of digitalization with the value proposition, understanding what you need to do and why,” Saenz says. In other words, fixating on the technology from the outset, whether it is internet of things tech, digital twin tech, or blockchain is like putting the cart before the horse. “Decide on the technology once your company knows which particular supply chain capabilities you need to enable, then the technology makes sense, the investment makes sense, and the return on investment from technology will make sense.”