Artificial Intelligence Is Changing Manufacturing

Over the last 50 years, some experts believe that parts manufacturing became complacent and failed to embrace innovation. That is why the parts manufacturing industry has struggled to keep pace with the fast pace of digital transformation today.

As Industry 4.0 technologies are growing, inefficient, and sluggish manufacturers are starting to feel the pinch. They have a choice of innovating or becoming outdated. This is being accentuated by the speed at which digital manufacturing is expanding and growing. (Supplychaindrive.com)

Today, digital manufacturing is starting to use technology to improve how parts are produced. The foundation of this industry-wide evolution is artificial intelligence (AI) and machine learning. Many believe that AI is one of the most disruptive technologies today.

AI and Machine Learning Worth $70 Billion in 2020

Some reports suggest that artificial intelligence and machine learning technologies are worth $70 billion in 2020. The machine learning industry could transform all types of business operations in all sectors. A report from NewVantage Partners in 2019 stated that 90% of C-suite executives believe investment in big data and AI is needed to develop and maintain competitive businesses. While there are many examples of AI benefits in many businesses, AI-equipped manufacturing businesses are reported as the most added value. (Mckinsey.com)

That is why Mckinsey expects the manufacturing industry to be one of the leaders in AI going forward, noting in its report that manufacturing is on the cusp of a revolution where AI applications will change end-to-end value chains with major shifts in demand. But much of the industry continues to struggle with getting started with cutting edge digitization. While most decision-makers think it is important for their companies to use AI, most have yet to get beyond pilot programs.

McKinsey also reports that most organizations in most sectors have begun to adopt AI in their businesses. A recent survey found that 47% of respondents said their companies are using at least one AI capability in their manufacturing processes, compared to only 20% of respondents in 2017. Another 30% said their companies were piloting AI. Still, McKinsey states there is much more potential to use AI across supply chains. Today only 21% of companies are using AI in multiple parts of their business. So far, investments in AI are still quite a small fraction of companies’ overall spending on technologies. More than 58% of respondents to the recent McKinsey survey said that only 1/10 of their firms’ digital budgets went toward AI.

The most common AI capabilities that have been deployed in manufacturing companies so far are:

  • Robotic process automation
  • Computer vision
  • Machine learning

For each of these, 20% said their firms had embedded these technologies into their manufacturing processes. Physical robotics and autonomous vehicles are less commonly deployed, largely because they are only useful for companies in specialties where there is a definite application for them.

Mentality Is the Major Issue Holding Back Manufacturing Companies

According to PwC, the biggest issue holding back manufacturing companies in the implementation of AI is their mentality. Many heavy manufacturing companies have a conventional engineering mindset that is averse to taking risks and is less about major internal process innovations.

That said, the benefits for manufacturing companies that use AI and machine learning in their daily processes could be massive. Machine learning, when it is used to assist human workers, can increase labor productivity by 40% and create major increases across value chains. For manufacturing, this is more than simply primary business functions, which is the actual manufacturing process. AI and machine learning also can benefit sourcing, maintenance and supply communications.

Machine Learning Can Be Used To Generate Quotes

A good example is using machine learning to improve the quote generation process. Usually, generating quotes can take up to three weeks to get approved. When an engineer is doing work on a product and needs a part for it, reducing the time for iteration is a big advantage for innovation. The period can be longer for many reasons, but human workers play a significant role.

Say you are making a part, and you choose three factories around the globe that can deliver the product you want. Your part is essential to the success of the product, but you are a smaller client, so the manufacturer does not give the production of your part priority. You may wait weeks for the quotes to be returned, and when they do, one is denied and the other two are too expensive.

Some companies may give you a higher price to encourage you to take your business elsewhere because they think making your part is not financially worth it. Others may think this is what the part is really worth. Whatever the case, the discrepancies are significant and are usually based on human evaluation of the situation.

Machine Learning Algorithm Can Greatly Increase Speed of Quote Process

When human bias is combined with the communication cycle that can take weeks, inefficiencies can pile up. But take the above example and use a machine learning algorithm to the process. Rather than the part quote getting printed out and left on a desk, when you upload the part and select your materials, the system would analyze the part right away and determine what it is worth and whether making it for a profit is possible.

With the major time and money-saving possibilities, the use of machine learning with manufacturing will become more important to keep businesses competitive and keep the sector growing.

While large manufacturers may be ok with operating behind the times, this will not work forever. The benefits of using AI and machine learning are major for manufacturing because they are embracing the latest innovation and technology. For those companies that choose not to, the scenario could resemble the one that faces Blockbuster and Netflix a decade ago. In the long run, if manufacturers really believe in giving engineers more power, the decision will be simple to make.