Supporting the backbone of innovation

Gavin: Where do you see manufacturing heading in the near to medium term?

Professor Goh: We may not see huge leaps and bounds, or exponential change in smart manufacturing in the next five to 10 years. Quantum computing or using hydrogen rather than oil, for instance, are paradigm changes, but those are the playbooks for very large companies with major R&D budgets or deep tech startups.

For most companies, we’ll see a lot of incremental improvements in operational processes that drive the backbone of innovation. There’ll be more integration, automation and dashboards, more use of analytical tools and AI to help with these manufacturing processes.

And more co-creation: the human-machine interface will become more important. I always ask my students whether they want to be managed by algorithms or be the ones managing the algorithms. Obviously, we hope they will be the ones defining the algorithms and managing the machines.

G: I hope so too! What's the responsibility and expectation on a major global company like Huawei, with its considerable annual investment in R&D, in this field of smart manufacturing?

P: First, segmentation by verticals and defining what solutions will look like within an industry and simplifying what the solution-set would be. Traditionally, Huawei sold hardware, but it’s going through a process of creating a services-and-solution approach. A solution approach has tight margins and is more value-added, which is helpful, but it also means IT companies need to better understand industry’s needs, so every solution would need to be customized that industry’s requirements. Second, as Huawei does a lot of in-house manufacturing, there is some level of control over pricing and an ability to package hardware and software solutions that are cost-effective.

G: Are we still in the early days of smart manufacturing, in terms of adoption?

P: There’s a big difference between large multinationals, which are highly automated, and small local enterprises, which, unfortunately, are in many ways starting off. Many do understand the importance of digitalization, but maybe don’t know where to start or it’s a cost issue. It can be hard to have enough economy of scale, and enough future efficiency cost savings, to justify the investment.

G: And is that changing, as technology costs come down?

P: Yes, with two major drivers. Covid generated a lot of digitalization as multinationals pushed for supply chain integration and visibility to really plan sourcing and shipping. That pushed their suppliers to comply and share information.

And there is the prevalence of software-as-a-service (SaaS), applications where you can essentially rent or pay per use. It's hosted on a cloud, and you can store your information and do transactional-based or subscription-based pricing for various types of applications, including manufacturing execution systems on the shop floor. That has helped adoption by small companies and probably also helped increase their digitalization as well. Because when service providers come on-site to understand your business and configure the software system to fit your processes, the companies themselves also learn a bit about digitalization and the use of IT, and discover it’s not as scary as they may think.

G: It’s interesting you use the word “scary.” How do you overcome those fears – whether they’re based on cultural, financial, regulatory, data security, or other reasons?

P: That depends on whose perspective you come from. For the management or owners, the concern will most likely be cost. If the ROI can be addressed, or if the multinational customer requires you to do it as a condition of doing business, then it is more likely to go ahead.

For the rank and file, the concern may be about job security.  One of the issues around transformation is always the fear that implementing technology could potentially take away my job.  So, that’s about how change managers work with existing staff to convince them that it is a positive change: taking away the tediousness of some of what they are doing, and letting them focus on higher value-added work, or giving them other opportunities.  Building trust and assurance from the company is needed.

And it’s really not the case that we can remove humans from all this. You have data, but you still need people to interpret what is happening, to troubleshoot when things go wrong – and actually, troubleshooting and interpretation are higher-level functions and require a fair bit of experience. Machine learning only works well within a box that we define.

But if something is exceptional or they haven't encountered and interpreted it before, then the machines still need to learn and that learning will have to come from human input. I think for the foreseeable future that human experience will be required. And in future, when people rely on these machines and fresh grads have not been through the operational and day-to-day functions? Even that could potentially be solved by technology. VR goggles, for example, could train them on the fly, showing them where to look for the issues, where to troubleshoot or where the manuals are that you can access on-site. Simulations can teach them. They may not have worked through the whole gamut of scenarios, but virtual training can speed them along the learning curve.

Cybersecurity concerns are harder to address for small companies – and even many large ones -- because they’re quite technical.  You have to trust that the cloud or proprietary systems you are using have the necessary tools in place to protect you.

G: There are obvious benefits for businesses in embracing smart manufacturing – efficiencies, profit, sharper analysis, a better service to customers. But amidst these jobs fears, does smart manufacturing offer a wider societal benefit, too?

P: Again, it depends on perspective. More developed economies don’t have many low-cost workers to rely on, so smart manufacturing helps alleviate resource issues. Even China has one of the most highly technical robotic implementations because it has a shortage of young workers. There are more issues for developing societies because a lot of low-cost workers may be affected. The biggest challenge is probably for less educated workers who are unable to use the tools of AI and ChatGPT for productivity gains. For the well-educated, it's a huge jump: better tools, generating ideas, better productivity. So, it has different impacts on different segments.

There are some potential benefits for society generally if you have more efficient, smarter cities, with sensors and cameras – you can plan in advance, get through traffic and airports quicker, etc. Privacy would be an issue, but different societies would have their own perceptions on that, and different willingness to share information. It’s why you see a lot of concerns about data sharing, and where data is being stored.

G: As a university professor who is Academic Director of a Masters Program in Industry 4.0 and teaching in it as well, do you think education systems are sufficiently prepared to give students the required skills?

P: Industry 4.0 is very broad because it encompasses various technologies that cut across the smart manufacturing space: robotics, IoT, 3D printing, and then on the software spectrum AI, data analytics, etc.

In the program, we want to create T-shaped individuals: a combination of breadth and depth. Breadth across different technologies, to understand how they work and enable you to manage implementation and projects; and also to manage expectations and evangelize within the company. And the depth is looking at one or two specializations, such as IoT or the digital supply chain, cybersecurity, 3D printing or robotics, etc. It’s a hybrid of both the business and technology worlds.

And there’s applied experiential learning as well, where they go and work with companies for six months to solve a company-sponsored project. Sometimes they say “Oh, it's a lot of work,” but that's the point: you learn in school, but you have to learn extra things and gain real experience outside school. Capstone projects are important, to bridge this industry gap. Careers are now all about continuous learning.