Manufacturers See Early AI Gains in 3 Areas

As more data is unlocked and more use cases yield results, we expect to see manufacturers moving along an AI/ML maturity curve—becoming bolder about how they apply AI and ML to their processes. This should lead to more AI/ML investments as organizations increasingly identify use cases. The upper end of the maturity curve is where we’ll see organizations consistently applying AI and ML across their shop floor, and where we will see broad employee adoption into their daily work.

Today’s AI/ML Use Cases

The use cases that some automakers and battery manufacturers are exploring now center on efficiency, quality and safety. For example, in the area of efficiency, AI and machine learning are improving production scheduling for a vehicle manufacturer that marries assemblies on its main production line. Without algorithms to learn how the process works and find ways to optimize it, the manufacturer must maintain a buffer of anywhere from four to eight hours to avoid line stoppage. That buffer also pushes up costs for logistics, warehousing and material movement through the shop.

Optimizing production timing: By unifying data to provide an ecosystem view that includes incoming assemblies and real-time line activity, Al/ML tools can analyze that data to optimize production timing. The AI system can then dynamically change the production schedule with suppliers, update receipt dates and locate safety stock if there’s a possibility of critical parts shortages. These capabilities can reduce the buffer from several hours to a few minutes, so assemblies are delivered just in time to go to the line. The result is lower costs and increased throughput.  Manufacturers we work with that are choosing to experiment with and implement initial AI / ML use cases are seeing immediate efficiency gains, as well as payback on their initial investments in 6-9 months post use-case implementation. 

Avoiding supply-chain slowdowns: Efficiency use cases could also apply to the supply chain. For example, EV battery manufacturers need to meet demand that will be six times greater in 2030 than it was in 2021, according to the International Energy Agency. Managing supply chain challenges to meet rising demand for battery minerals is the “front and center” challenge for the EV market, according to the director of the U.S. Department of Energy’s Argonne National Laboratory. AI tools that help manufacturers source inputs based on real-time availability could avoid slowdowns caused by shortages.

Improving battery chemistry: AI and ML are also starting to play a critical role in optimizing EV battery quality. For example, some digitally native gigafactories are combining extensive automation with AI-powered data analytics to optimize the way they mix battery inputs. By tracking the chemistry and results for different mixtures, these gigafactories can identify the best-performing combinations of minerals and other inputs to improve battery quality and consistency, avoid waste and control costs.

Making operations safer: The volatility of chemical battery components can pose safety risks on the shop floor, especially for manufacturers that are not native to the chemical manufacturing space. AI has the potential to increase the safety of working with and storing the chemicals used in EV batteries by identifying best practices and analyzing real-time production and storage data.

Putting Together a Deployment Plan

While the use cases are in motion, it’s time to think more broadly about deploying AI/ML in vehicles and EV battery manufacturing.

Understand data obstacles: Because data is the fuel for AI learning models and the material that machine learning algorithms work on, battery and vehicle manufacturers who want to start building AI/ML use cases can start by understanding their existing data architecture. It’s critical to understand how accessible production data is, what barriers may impede access and how to remove those barriers.

Prioritize use cases. After data, the next step is to inventory use cases across the production process and assess the potential value that AI and ML can create for each. The idea is to prioritize the use cases that offer the quickest wins first, to get the innovation flywheel going and start building muscle memory for eventual AI implementation across the organization. As the initial use cases deliver results, it becomes possible to scale the architecture to apply to more use cases and create more value. Change management plays a critical role here as initial use cases are launched. Organizations need to ensure that their shopfloor employees are not viewing these new solutions as threats to their jobs, but rather as an opportunity to shift toward more value-added work.

Lead culture change. Along the way, organization must also shift its culture beyond its legacy improvement practices to embrace data and AI as the new drivers of optimization and value creation. Leaders and workers must be on board for this mindset evolution to succeed, which means everyone in the organization needs to be engaged in the process from the outset and contributing to the continuous ideation that initial use cases trigger.

Moving from traditional process-improvement techniques to AI-backed optimization programs is the key to getting past the plateau many auto industry manufacturers are on now. Building the institutional muscle memory for this change requires starting small and working methodically toward new possibilities in efficiency, quality, safety, and other kinds of value creation.