8 use cases for generative AI in manufacturing

1. Product design and optimization

GenAI can speed up product design by automatically creating numerous design alternatives based on designers' goals and constraints. AI art uses a similar approach.

In the manufacturing design process, engineers submit requirements and parameters into the generator, such as materials, cost limits, weight and required strength. While human engineers might stick with familiar patterns, AI-generated designs can create a wider range.

Manufacturers should be sure to be aware that their company is staying aware about the latest legislation on AI creations and patents. In addition, engineers must critically vet all AI-generated designs to make sure they can be manufactured effectively. These designs must also meet safety or compliance standards, just like human designs. Engineers still have an important role in refining or adjusting the AI's output to meet real-world constraints.

2. Quality control and defect detection

In manufacturing QA, GenAI can help detect product defects and is often able to do so more reliably and earlier than traditional methods.

Engineers train AI models using data sets of images and sensor readings from both high-quality and defective products. The models then learn to distinguish many kinds of flaws. In situations in which defects are rare but critical, the system could potentially create synthetic examples of rare problems to improve training and process optimization.

On the production line, computer vision systems can examine products in real time and spot small defects or anomalies that human inspectors might miss, such as hairline cracks.

3. Predictive maintenance and equipment monitoring

Predictive maintenance is a potential use case for GenAI in manufacturing. Machines send real-time sensor data that typically includes readings for vibration, temperature, pressure and noise. Any change in these readings could indicate a problem. These anomalies might be too nuanced for traditional rule-based equipment monitoring systems to catch, but AI can recognize the early warning signs.

Operators can use these insights to schedule maintenance before a breakdown occurs, replacing the part or tuning the machine during planned downtime rather than operators needing to react to an unexpected failure. A real-world example of this is modern mechanics accessing a vehicle's sensor data to figure out whether preventative work can eliminate issues in the future, such as faulty brakes.

4. Supply chain optimization and demand forecasting

GenAI can potentially help improve supply chain management through demand forecasts and by improving logistics in delivery.

Large data sets, such as historical sales, market trends, seasonal patterns, macroeconomic indicators and weather, help AI predict product demand more accurately than traditional systems. Manufacturing leaders can use that product demand data to figure out how much their company will need to produce to maintain optimal inventory levels. Supplier performance and warehouse ability limits data also help inventory recommendations be as accurate as possible.

Beyond forecasting supply chain issues, GenAI can help create more efficient routes and schedules by analyzing factors such as shipping data, traffic and delivery requirements. An AI system might be trained to generate a trucking route that accounts for real-time traffic or suggest reordering shipments on a schedule that will minimize fuel usage and transit times.

An AI model is also useful for "what-if" scenarios, as it can model the effects of rare but impactful disruptions, such as a bridge closure or geopolitical disruptions.

5. Process automation and efficiency improvement

A promising AI application is creating digital twins of manufacturing processes. Digital twins are virtual simulations of factory equipment, production lines or entire facilities. Manufacturers can monitor operations virtually by feeding real-time sensor data into these AI-driven systems and running simulations to optimize them.

GenAI can simulate thousands of variations of a production schedule or an assembly line configuration to find the most efficient plan. The model can also suggest the optimal ways to allocate resources on the shop floor, design better facility layouts, and generate improved machine toolpaths for faster production.

These recommendations help engineers and managers refine their processes, so operations are more efficient and produce less waste. Once a digital twin is up and running, AI systems switch from advisors to orchestrators, automatically adjusting parameters to avoid bottlenecks and stoppages.

6. Customization and mass personalization

Mass personalization is a growing trend in manufacturing, and GenAI can be a key component for producing customized products at scale. In the past, bespoke styles or sizes needed expensive and slow handcrafting. Now, GenAI can help manufacturers quickly produce and adapt product designs or specifications to meet individual customer requirements.

In practice, an AI-driven design tool receives a customer's input, such as a custom texture for the sole of a running shoe. The system then creates the design code for the manufacturing robots as well as instructions for human quality inspectors.

7. Workforce training and knowledge transfer

Manufacturers spend significant money on training staff in complex processes. GenAI can help create personalized training that's matched to each worker's role, skill level and learning pace, which can be more beneficial than one-size-fits-all training programs.

A GenAI model can create training materials, administer tests and evaluations, and then analyze data, such as an employee's performance metrics in production. The content can be delivered in the format that would best suit the learning objectives, which might be text, video, or simulations.

For example, a new hire in assembly might complete an AI-generated interactive tutorial that focuses on the specific assembly steps and machine interfaces they will use for their work. Meanwhile, a maintenance technician might receive a different set of AI-created exercises that target the machinery they oversee.

8. Sustainable manufacturing

Sustainability is a growing priority in manufacturing, and GenAI can increasingly help companies make their production more environmentally friendly. AI can potentially find patterns in the ways that machines consume power and then suggest optimizations, such as adjusting load distribution or scheduling energy-intensive tasks at off-peak times.

Manufacturing leaders can also run simulations to figure out the settings or process configurations that would minimize energy usage while maintaining efficiency. For example, GenAI might simulate a manufacturing line's behavior at various conveyor speeds or analyze oven temperatures to find a combination of temperatures that use the least energy per unit produced without affecting quality.

Similarly, GenAI can help reduce waste and materials by optimizing product designs and process parameters so manufacturers use raw materials more efficiently. For example, AI could reduce waste and material requirements for injection molding or 3D printing by tweaking a design.

Donald Farmer is a data strategist with 30-plus years of experience, including as a product team leader at Microsoft and Qlik. He advises global clients on data, analytics, AI and innovation strategy, with expertise spanning from tech giants to startups. He lives in an experimental woodland home near Seattle.