May 7, 2019 Striving Towards Self-Driving Supply Chains

They need to be able to make sense of structure and unstructured data from various sources, have current info about all their suppliers, track all their products and inventory pools, and ensure on-time delivery to all customer orders, etc. Compounded by customers’ fickle nature and rapidly shifting loyalty, organizations must also promptly and accurately address customer queries – or risk losing business.

To achieve this, organizations need to be able to take advantage of machine learning (ML) and other artificial intelligence (AI) technologies, and perform not only predictive but also prescriptive analyses. With these technologies in hand, the goal is to achieve continuous and self-driving supply chains – that will be propelled by digital technologies with minimal human intervention.

Evolving Supply Chain Challenges (and Appropriate Responses to Them)

Today’s supply chain is no longer your father’s supply chain, to put it bluntly. Empowered and fickle consumers expect to buy everything anywhere, have it delivered anywhere, and return it anywhere—no questions asked. Compounding these high expectations are increased global competition, volatile market demand and geopolitical situations, ever-shorter product life cycles, fast fashion-quick manufacturing trends, and consumer expectations for responsibly sourced materials.

Decades ago, companies could do well with silos of raw data, inside-out processes, limited visibility, inaccurate forecasts, and other inadequate assumptions and subpar practices. Such retrospective management practices with descriptive analytics at best, which can only tell you what happened well after the fact, are a losing proposition these days.

Meeting the challenge of managing a modern supply chain, for any complex organization, requires near real-time analysis of customer orders, tracking of complex and diverse pools of inventory, and up-to-date knowledge of suppliers, and any other issues around the world. As much of this information still lives across data silos in legacy enterprise systems, it is extremely difficult for any large enterprise to dynamically assess its production stocks, deliver products on schedule, and respond to customer demands with agility.

These days, large enterprises with complex business operations and processes need to be able to react to and handle the outside-in processes within holistic value networks rather than in-departmental silos (with inside-out processes only). They must also be able to handle both the traditional structured data (i.e., traditional numerical and textual data tables) and unstructured data (i.e., weather info, social signals and sentiments, etc.) to be able to sense events and promptly produce an intelligent response. Also, these businesses need astute autonomous and localized responses, as the one-size-fits-all standardized global policies no longer satisfy the needs of the differing markets.

Some forward-thinking and innovative manufacturing companies have taken steps to achieve harmonized end-to-end real-time visibility and enriched external data. Essentially, they need to build a cognitive data model. A cognitive time series model for demand sensing uses machine learning and other artificial intelligence (AI) tools and methods to produce two types of forward-looking analytics. One is predictive analytics to provide answers to questions such as “What will happen?” and the other is prescriptive analytics to answer the question “What should we do about it?” and other similar questions.

Over the longer term, the goal is to achieve a closed-loop and continuous (autonomous and/or self-driving, if you will) supply chain. In that ideal (nirvana) scenario, machines can make some operationalized decisions, perhaps with some minor human supervision and input.

Click here to read the full Industry Report on the TEC website.

About Technology Evaluation Centers (TEC)
Technology Evaluation Centers (TEC) is a global consulting and advisory firm, helping organizations select the best enterprise software solution for their needs. TEC reduces the time, cost, and risk associated with enterprise software selection with its advanced decision-making process and support application, software selection experts, and extensive resources.

Over 3.5 million subscribers leverage TEC’s industry-leading research and detailed information on more than 1,000 leading software solutions across all major application areas. For more information, please visit www.technologyevaluation.com.

Subscribe to Industry Today