Unlocking the Multi-Billion Dollar Potential of Generative AI in Manufacturing
The manufacturing industry is on the cusp of a significant transformation, and generative AI is poised to be crucial. In the first article of this series, we explore where the business value lies and what potential and roadblocks lie ahead of generative AI in manufacturing.
The Imperative for Digital Transformation in Manufacturing
Manufacturing has long been a focal point in Industry 4.0 and broader discussions about digital transformation. However, the sector's inherent complexities in production modes, practices, and scalability pose opportunities and challenges for embracing new technologies.
The urgent need for technology adoption is underscored by obstacles such as the shortage of in-house AI expertise and the risks associated with disrupting production cycles. Moreover, recent events, like the Covid-19 pandemic, have impacted manufacturing supply chains. It has also presented a new opportunity, with PwC reporting three-quarters of 5,000 surveyed CEOs are optimistic about economic recovery and seeing a significant acceleration of digital transformation efforts.
Generative AI can serve as a new booster for industrial digital transformation. More people can taste the technology firsthand with individual experts accessing tools like ChatGPT. Most importantly, people without technology expertise are now empowered to demand new, better solutions to help them with their daily tasks at work. Let’s investigate what this means in manufacturing.
Harvesting Value from Data with Generative AI
Generative AI will impact all industry sectors. The total impact ranges in trillions of dollars, with a McKinsey estimate ranging from 2.6 to 4.4 trillion USD. The estimated value exceeds 100 billion in manufacturing and 200 billion in supply chains. Even though not as lucrative as other sectors, this underscores the massive opportunity of generative AI. Embedded here is a chance to tackle existing challenges in the field, create new business value, and generate fresh investment interest all at once.
However, the correct AI solutions for strategic business needs require data. Manufacturers collect plenty of data, including machine operation data, tool and material usage information, shipping and order logs, and production data like downtime, output, and maintenance. Yet, universal manufacturing data on which AI could be trained is lacking.
A common issue is that data available is local or otherwise specific because managers are most often encouraged to focus on matters affecting their site. Moreover, data collection itself can be burdened with the high costs of setting up data collection, needing to sync incompatible systems, or keeping numerous systems up to date.
Such issues are standard obstacles in an industrial data strategy. Still, generative AI is possibly the first solution that can address this entirely and be the foundation for new production opportunities. Such identified by McKinsey in this area include using generative AI for data discovery, ingestion, storage, processing, access, consumption, governance, and interpretation.
Existing Challenges: Empowering Workers with Generative AI
Once manufacturing data is primed for generative AI, the foundations are set for coordinated, wide-scale implementation of worker applications. Generative AI will primarily affect workflows and ways of working. Planning to enable improved human-machine collaboration with the new technology is paramount for unlocking its value.
One of the most anticipated advantages already links data and humans together inexorably. With an aging workforce, retiring specialists with decades of experience risk retiring plenty of their knowledge with them. Generative AI can address this, and initial use cases are already available. For example, a study conducted at a Fortune 500 company customer service division found that generative AI applications help increase issue resolution by 14% an hour, with junior staff members benefitting from the tacit knowledge of senior workers.
Such applications can help retain essential knowledge in the factory. Moreover, it is not limited to knowledge in individual minds, but other more or less unstructured sources too. Technicians' notes and visual imagery can be structured in such a way as to retain knowledge and improve operational efficiency. Such an example has been demonstrated by Siemens and Microsoft with images and video captured on a factory floor to train generative AI vision models.
Solutions placed at the end of the production line can help, too. Combining generative AI with visual sensors can support quality control and even synthesize written data for human operators to read. These solutions will save time by reducing recalls and rework.
New Solutions for Machine Interaction and Supply Chains
Generative AI can also enable entirely new features on the factory floor. Deploying tools for specific manufacturing contexts gives operators and maintenance crews new ways to increase their productivity and welfare in a production environment.
One exciting use case is the potential to have a database of documentation that can be queried quickly. With quick recall of specific instructions or engineering notes, generative AI can enable a new level of continuous operation where stoppages are rare.
Humans are not the only ones to benefit from generative AI: enhancing manufacturing robotics with AI assistants can achieve a new level of collaboration between humans and machines. Manufacturers like BMW and Ford are already deploying such devices for assembly line procedures like gluing and welding.
AI's advantage in this case is that machines don’t have to be separated from humans with cages and other protective gear; their programming can consider human operators and watch out for them. With generative AI, such work could help humans and machines interact with voice or hand signals when needed.
Outside of the factory, generative AI can assist in planning and optimizing supply chains. Uncertainty has been widespread recently, with the Covid-pandemic being an absolute high point. In such situations, generative AI can be tasked with suitable optimization routines, like automatic alternative sourcing operations. With insights from AI systems, manufacturers can get ahead of potential disruptions before they even take place.
In Our Next Blog: Design as Part of R&D and Product Development
At the beginning of any production process is the product design itself. With lots of time and resources dedicated to creating products that are easy, cheap, and fast to manufacture, generative AI once again promises to open a new frontier where historical data and further analyses can break open a piñata of reclaimed productivity.
While AI and machine learning have already been deployed in this area, generative AI is the key to creating precise designs with contextual goals as priorities. For example, the availability of certain materials at a specific manufacturing site could become a factor that defines a much more sustainable design in the future.
Of course, design is not an exclusive feature of manufacturing. This post is the first in our series covering the themes under Generative AI for Industry. In our next series post, we will explore generative product design more as part of our exploration into R&D and Product Development.
Want to work with the Foundry companies to develop the next great manufacturing generative AI solutions? Read more about our specific needs here and submit your solution!