AI and a New Era of Manufacturing? Believe the Hype

Q&A With McKinsey & Company’s Aaron Aboagye

Based on your work in Michigan and elsewhere across the U.S., what trends do you see in terms of how AI and generative AI are shaping the future of manufacturing?

AI and Generative AI (GenAI) may end up being innovations where the reality actually lives up to the hype – or even exceeds it. In our research with the World/Economic Forum, we’ve focused on socalled “Lighthouse” sites globally that have proven at-scale AI implementation and are seeing incredible gains (all 24 of them have a mature GenAI pilots underway). Across the global economy, GenAI is projected to add between $2.6 trillion and $4.4 trillion in annual value, with up to 25% of this value coming from manufacturing and supply chain productivity improvements. By 2030, we’ve estimated that up to 30% of current hours worked could be automated.

This automation is driven by emerging AIenabled technologies in robotics, predictive maintenance, dynamic scheduling, quality resolution, and new product design among others. We’re at dawn of a new era, which will require thoughtful management.

McKinsey works with many CEOs, CIOs, COOs, and other leaders on AI Implementation—what are some key challenges, concerns, or misconceptions that you hear the most?

As with any big change, the tricky part is how to balance opportunity and risk –and there’s plenty of both to go around. With GenAI, large employers have likely made some significant investments during year one, with with a range of interesting pilots underway. Our research shows the big question for leaders is: How do you avoid getting stuck in pilot purgatory and achieve truly transformational impact? This frequently involves conversations about better defining the strategy, data infrastructure, acquiring and retaining the right talent, and capability-building (learning new technology as well as new roles). The part people sometimes forget is that big transformations are about technology AND operations – you need to nail them both.

What specific AI applications have shown the most promise in industrial processing plants, that may be applicable for Michigan’s manufacturers?

McKinsey’s research highlights a wide range of promising AI applications. Looking across the factory, it’s everything from creating AI command centers for end-to-end operational automation to using GenAI for supplier risk forecasting, and providing support to technicians. We’re also seeing promising applications for real-time process parameter optimization, enhancing production quality and efficiency, and predictive maintenance that helps reduce downtime and costs. The potential and speed of this transformation can be daunting for leaders to navigate (while still pushing on everything else). Our role is to help identify and synthesize the best practices we’re seeing around the world and help clients find the right solutions.

What are some of the workforce implications that leaders should have in mind related to AI implementation?

One interesting nuance with GenAI adoption is that given the excitement around the technology and its ease of use (it speaks English!), it’s often the employees themselves who are jumping in first, while employers have taken longer. Our Global AI survey, for example, found that 91% of respondents (employees) were already using GenAI for work, whereas only 13% of companies had implemented multiple use cases. So where to begin? Companies often start by focusing on domain-based transformations (e.g., in software development, marketing, customer service) to integrate GenAI into value-creating workflows and processes. The key is that it is part of a holistic strategy. One caution is that the people who are most adept at experimenting with GenAI and creating use cases are often the biggest flight risks.

What are the potential risks associated with generative AI in manufacturing, and how can companies effectively manage these risks?

Similar to other new technologies and big transformations, you can break the risk into a few broad categories relating to the technology itself, people and organizational challenges, and a broader legal/compliance framework. In the first category, people are now familiar with the concept of hallucination (inaccurate data) in GenAI. You need policies to mitigate that – making clear decisions about when you need humans “in” the loop, versus “on” it. In terms of the broader organization, AI and automation are retooling how factories operate, including what people and skills are required. With the “Lighthouse sites” mentioned earlier, those facilities create an average of 25 new digital roles –about half focused on data and AI – per 1,000 factory workers. On the compliance front, it is often questions about data privacy and security, intellectual property, and broader regulatory compliance for a field that is still emerging.

Aaron Aboagye is Partner at McKinsey & Company and is a board member of the Detroit Regional Chamber.