Manufacturing is often seen as a ‘poster child’ for technology such as artificial intelligence (AI)—but is manufacturing a smart industry on the whole?
Globally, the image of robots making things and the broad uptake of AI is not really a reflection of our industry. The reality is most manufacturers are small- to medium-sized enterprises (SMEs) with staffs smaller than 500 employees and less than $200 million in turnover. Moreover, their uptake of AI and smart technologies is often very low.
At TilliT, we’ve found that up to 90% of SME manufacturers are still capturing information on paper (such as production metrics, maintenance or quality checks) making the goal of ‘lights out manufacturing’ a distant prospect.
On top of that, 75% of the global manufacturing asset base is disconnected—there's currently no AI, no cloud analytics, and no smart Internet of Things (IoT) sensors analyzing the behavior of this equipment. Essentially, they just have a machine that is isolated carrying out production processes without anything smart happening. So, these companies simply aren’t set up with an appropriate technology foundation, and they don’t see themselves reflected in movement towards these advanced approaches.
This is where the concept of smart manufacturing can help bridge that gap.
What do we mean by smart manufacturing?
Smart manufacturing is an application or system that uses intelligence to integrate factory processes. It includes the use of complex computer-driven optimization decisions that humans may typically make.
These two key words—integration and intelligence—define the essence of smart manufacturing and are applied to physical processes, equipment, people, and even behaviors.
If the objective is to increase profit, reduce risk, or improve efficiency, then you need to accept that innovation is the key to success. The traditional way to innovate in manufacturing is to buy new equipment and automate it. Even today, automation and most robotics in manufacturing plants are run by simple programmable logic without the use of any AI. Automation is still a way to achieve great results, and when technologies such as AI and machine learning are applied, it can super charge those efficiencies.
So, why aren’t we seeing more AI or machine learning in manufacturing?
Industry fear of failure means less adoption of these initiatives
Gartner predicted 85% of AI initiatives would fail within manufacturing, so it is understandable companies hesitate to adopt these technologies.
Failure occurs due to many factors, including lack of investment in the change management process, a lack of preparedness or understanding in the process, or lack of internal knowledge and resources. Importantly, there’s typically a lack of verified, accurate real-time data to build AI models on.
We think it’s important that advanced techniques like AI are combined with the understanding that people involvement in the process is a reality for years to come. In most manufacturing industries, people are still likely to be a cost effective and efficient resource.
If you’re unable to support the operator with AI, prescribe what tasks need to be done, and track the execution of those processes it’s difficult to learn from and add intelligence. Basically, no matter how good the AI is, if the equipment and people do not act on the recommendation then it is wasted.
Digital manufacturing in practice
In a smart manufacturing environment, one of the first requirements is to digitalize processes—removing paper, and moving to connected, digital and automated workflow driven operations.
A well-planned connected workflow can capture tribal knowledge, detailing cause and effect relationships and making operator tasks more efficient. The operator can do more, and the manager knows exactly what’s happening on the factory floor with a view of both equipment and personnel.
We’ve seen examples of this with two recent customers:
An electrical component manufacturer TilliT we worked with got a baseline digital manufacturing platform in place to capture information that was previously paper based in its business.
Using a prescriptive algorithm, we injected a production scheduling solution, to start improving the thinking about business operations, including how they could:
- Optimize injection molding
- Reduce changeovers
- Smooth staff allocation / shortages
- Predict future production rates
For a kitchen sink manufacturer, replacing their existing press machine—which is close to 40 years old and has no automation—impacts production.
TilliT’s approach was to retrofit IoT sensors to the machine, capturing rates, stoppage and downtime reasons. Using this information to then feed prescriptive AI algorithms and scheduling improvements, our client achieved tangible results:
- Increased throughput by 2%
- Reduced work in progress by 25%
- Optimized inventory
- Reduced material wastage.
For these clients, people were critical enablers in achieving these results.
Rather than excluding personnel from this future vision, the manufacturing industry should be embracing people involvement, engaging them with advanced capabilities and techniques and delivering a different type of AI—Augmented Individuals.
James Balzary is the CEO and co-founder of TilliT, a SAGE Group product. SAGE is a certified member of the Control System Integrators Association (CSIA). For more information about SAGE Group, visit its profile on the Industrial Automation Exchange.
