By Nikhil Chauhan
The Industry IoT Council is an Industrial focused Executive Community comprised of business executives, technical experts, and academics representing a diverse array of global industries. The group convenes periodically to share insights regarding the impact and potential benefits of incorporating new and emerging technologies in day-to-day business operations. Following is a recap of a session focusing on the topic of An Adaptive Future.
The Adaptive Future is here. Are industries ready to harness its power?
Data-driven, adaptive technologies are beginning to transform business processes across a broad spectrum of industries. While companies are certainly cognizant of the benefits that AI, analytics, and automation can deliver, many are not fully prepared to get on board the Adaptive Future where they can integrate these powerful, adaptive technologies into their manufacturing and supply chain operations. What are the hurdles they are facing and where have they seen success? Although still in its early stages, Industry 4.0 is inevitable and the market for the technologies driving transformation is growing rapidly.
Data, data everywhere. But how to make use of it?
As noted by several during the April Industry IoT Council discussion, leveraging machine learning and analytics may require building an entirely new digital platform in less technically advanced industrial environments. But regardless of how savvy organizations are on the technology front, they all need to deal with the realization that data is the lifeblood of adaptive technologies. So, it’s no surprise that the major benefits — as well many challenges — to be had by integrating data-driven applications into supply chain and logistics workflows revolve around the way data is ingested, stored, processed and presented. With always-on IoT devices, smart machines, IT systems, and various applications streaming and using massive amounts of data in real time, companies are being inundated. To minimize the risk of overwhelming users, new ways of “feeding the system”, leveraging data and analytics at the edge; and ingesting small, wide, unstructured and structured data are being devised. These dimensions require modern approaches to analytics and platforms.
Planners need the right data in the right format at the right time to make good decisions. To make data actionable and useful for decision-makers and avoid “analysis paralysis,” analysts need to learn how to translate technical vagaries into a language that decision-makers can actually use. Nobody reads algorithms. There’s a learning curve on both ends — technical and managerial — that can be difficult to navigate. Multiple training sessions may be required. As noted by one participant, bridging the gap between Big Data analysts and operations (in this instance the actual physical movement of electronic components through the supply chain) can be rather challenging. “They live in two separate worlds.”
A question of trust
Data overload is a key issue. Knowing what can be trusted is equally critical. Security is a major concern, with malware that can corrupt data and ransomware attacks that can bring operations to an abrupt and costly halt. The risks multiply when sharing data with external parties such as vendors, partners, and customers. To ensure data integrity, centralized, corporate control is advised, with little or no local autonomy granted to extraneous business units, remote locations, or third parties. On the corporate front, the willingness to trust data-driven decision-making is eroded by fear of the magic “black box” — with data going in, and answers are spit out. Do we really understand what’s going in the box and can we really trust the recommendations coming out?
Who’s in charge?
Sorting out the organizational dynamics impacting analytical projects is another hurdle. Who decides what data to collect and who gets to use it?. In the case of our session panelists, they mentioned it has typically been business intelligence, marketing, operations, and supply chain managers who have had oversight and were charged with demonstrating a project’s value and getting buy-in from corporate executives to put up the money. Creating a solid business case is critical. Several panelists recommended starting out with limited-scope, pilot programs designed to demonstrate the productivity and/or cost-cutting advantages that predictive analytics, AI, and robotics could potentially deliver.
The Industry IoT Council Responds - a survey from the session
What’s your top priority for leveraging data?
• Increase Profits – 15%
• Increase Productivity – 54%
• Both Equally – 31%
Not just technology
While digitalization does play a central role in modern supply chain management, it’s not the only factor. In the past year, travel delays and quarantines in the wake of COVID-19, the Suez Canal tanker episode, and the Texas deep freeze have all created global supply chain disruptions and logistical nightmares. And then there is the human element. Supply chain managers note that supply planning is difficult to get right because customers typically exaggerate (or underestimate) the quantities of product they will need, and contract agreements and POs often have to be signed as much as 18 months in advance.
During the Adaptive Future session, Industry IoT Council members cited a number of applications they have implemented to improve operational workflows:
• Multivariate data analysis — Delivering key metrics in real time throughout product development processes to ensure adequate ingredient sourcing and improve product consistency and quality.
• Demand planning — Crunching historical data to forecast future demand across complex, multi-vendor supply chains.
• Predictive analytics — Building models charting out optimal global distribution routes for spare parts.
• Process automation — Integrating information-rich, process-based data in SAP to streamline material flows across a global supply chain.
• RPA — Piloting industrial robotic trials to highlight productivity gains and — equally important — demonstrate that robotics can, “help you work smarter, make your life easier, and not take your job.”
• Digital Twins for Predictive Maintenance — Pairing data analytics and smart goggles in industrial factory environments to enable visual overlays (red, yellow, green) that put the spotlight on machines in need of imminent repair.
• Supply chain planning — Using photo scanning and machine learning to recognize thousands of spare parts, create future demand models, and automatically sort and route the parts to optimal storage locations.
• Safety — Collecting data from electrical transformers to enable predictive maintenance across the power grid.
• Root Cause Analysis — Correlating historical oil and gas prices with external factors to understand the causes of price fluctuations and create profiles for future sales and pricing expectations.
• Power Allocation — Exchanging data between various suppliers (e.g., hydroelectric, solar, nuclear) and creating models of future energy needs to ensure appropriate power supplies are delivered when needed, but only enough, so none is wasted.
Can’t let the machines run wild
An Adaptive Future fueled by automated, AI decision making is still very much in the early stages. Nevertheless, PwC analysts predict that by 2030 up to 70 percent of GDP growth worldwide will spring from AI and automated robotic technologies. Industries will need to carefully support ethical AI and consider how much authority machines will ultimately be given moving forward. Currently, the Industry IoT council members indicated their most common usage around AI and analytics related to creating systems of “augmented intelligence” with lots of human oversight and intervention in the loop. While there are a lot of new trails to blaze, industries will need to carefully consider how much authority machines will ultimately be given moving forward.