Artificial intelligence (AI) is gaining a lot of attention and the interest has accelerated like other areas of technology as a result of the pandemic. If you are a mid-size manufacturer, your attitude might be that AI is for the big boys, not guys like us. However, the world has changed significantly in 2020 and manufacturers need to adapt to new realities. I’ll discuss AI in more detail below but in the meantime, here are some points for you to consider.
- AI has the potential to transform performance and quality across the spectrum of manufacturing operations. As more manufacturers adopt Industry 4.0 technologies, like the Internet of Things (IoT), AI will be needed to help analyze data for value-adding tasks such as predictive maintenance or performance optimization.
- Analyst groups like IDC and Gartner forecast that spending on AI will grow significantly in the next three years (e.g., Spending on AI in 2023 will be more than two and a half times the spending level of 2019 (IDC Worldwide Artificial Intelligence Spending Guide).
If you now believe that AI needs at least some serious consideration, let’s look at the emerging role of AI in manufacturing.
Benefits and use cases
There are definite measurable benefits to using AI in a manufacturing business.
– automating a process to augment human ability, e.g., running back-end operations like managing a production network;
– optimize the efficiency of a process, e.g., identifying defective products;
– enhance the ability of people to accomplish tasks or enable them to do something they typically could not, e.g., improving compliance.
There are several use cases that a manufacturer could consider when kick-starting their AI journey.
Predictive maintenance
Using AI in conjunction with data from IoT sensors and other sources enables better prediction and avoidance of machine failure. Capital equipment productivity can be increased up to 20% and maintenance costs may be reduced by up to 10%.
Yield enhancement
Manufacturing that is supported by AI can see decreased scrap rates and testing costs by linking many variables across machinery groups and sub-processes.
Quality testing
AI-based image recognition can significantly increase the detection of defects compared to human inspection. Because these AI systems can learn continuously, their performance improves over time.
Supply chain management
Using AI can improve forecasting accuracy at an increased SKU granularity with a reduction in errors between 20% and 50%, and inventory reductions of 20% to 50%. AI does this by analyzing and learning from various data sources such as warehousing and inventory data in an ERP system as well as from external sources like social media information.
R & D
AI can improve product design by iteratively testing and learning which can optimize designs and suggest solutions that may appear unconventional to the human mind.
Shop floor operations
Combining real-time monitoring (using sensors) with AI can optimize shop floor operations, providing insights into machine-level loads and production schedule performance. AI can analyze large amounts of data coming from sensors far more effectively than humans and make recommendations in near real-time to assist human decision making.
Health and safety
AI can be used to get a better understanding of risk factors on the shop floor and can help make operations safer.
Implementing AI
Unlike other business systems that usually get implemented all at once, because AI is so new and people need to learn about its issues, AI projects are often implemented as small-scale prototypes in live environments. The first step is to get the AI prototype processing data in real-time from the shop floor or warehouse. To automate the collection of real-time data from live production operations will require integration to a manufacturing operation management or manufacturing execution system; this can be assisted by data from IoT sources.
Scaling up beyond the prototype phase is not a challenge to be under-estimated. Performance needs to be continuously monitored for quality and reliability, and for value generated. However, once the AI solution is proven and ready, the application can be deployed and made available across multiple sites. As you scale up, more value is extracted from AI as it gets applied, and learns, across divisions and geographies.
Procedures for auditing and testing should also be adopted to ensure AI risks are understood and addressed throughout the organization.
Combining AI with ERP
When it comes to AI applications, about 50% of businesses are buying as opposed to building their own (Deloitte: the State of AI in the Enterprise). Since you are going to need data from an enterprise system as both input and output for an AI system, it makes sense to consider what your ERP system can offer.
AI as a first-class citizen of your ERP
The AI solution should be a first-class citizen of your ERP, not a third-party add-on. Make sure that AI is an integral part of your ERP.
Security risks
Add-ons become security risks. A third party AI solution creates the problem of handling a large amount of data, with the security issues that involves. It is better to integrate it into one system so that the data is not put at risk. By having the AI system integrated with your ERP you also reduce your time to deliver a full solution.
The solution should be for everyone
Your ERP should enable anyone to engage with AI, without them needing to be data scientists. Unless you understand the data you’re being shown, there is no value in AI for you. You need to make sure existing automated report writers can read the AI reports. Look for an ERP solution that has a report-writing tool that will allow you to bring AI information to the user interface.
Getting AI right
If you are still wondering whether to go the AI route or not, remember that if you do not change the way you operate, the tools you use, and the degree of automation you choose, your industry and competitors will do and your customers will move on.
AI adoption will alter the composition of the workforce by lowering the need for manual activities in production processes. But as existing jobs are eliminated, new jobs with skills that complement AI will arise.
Finally, for manufacturers to get success from AI, they will need to see revenue from it, their AI strategy will be tied to corporate strategy, their organization re-aligned, and ensure that AI can be deployed across the enterprise.
