Intelligent manufacturing: The Canadian story

The Canadian manufacturing industry is well aware of the importance of embracing artificial intelligence (AI). AI has been a staple in digital transformation technologies for several decades now. As it has evolved, it has opened new pathways to transform how the industry functions and grows.

So, where does Canadian manufacturing stand today? In the inaugural ALL in AI event that took place in Montreal in September, Isabelle Hudon, president of the Business Development Bank of Canada, shared that the level of adoption and maturity is quite low, according to currently available data.

“In Canada, we do lack adoption of technology to start with when we do compare our country to other countries. Digital adoption is pretty low, and way too low if we want to be better and more productive. It’s tough to ask SMEs to leapfrog over digital adoption and go right to AI adoption. So, it’s quite a journey. I do think that all of us, researchers, academic people, entrepreneurs and leaders in the field, have to develop a more tangible narrative around the benefits of AI. Also, we do talk about AI too often in a very complex way and show applications that are one, very expensive, and two, quite inaccessible for SMEs,” she said.

Hudon added that industry leaders should come back to a narrative that makes the definitions and benefits of AI simpler. This, she said, would encourage adoption and increase the speed of adoption. She shared that manufacturers that do adopt AI are able to see an impressive difference in their growth and efforts in scaling up their businesses.

With so many advanced technologies and choices, there might be a certain level of hesitance from manufacturers in investing in AI without understanding what it can help them accomplish.

Hudon explained that to close the gap in adoption, the focus should be on education.

“I do believe that it’s non-negotiable to train entrepreneurs and the business community on the many benefits that AI can bring into one business and into their businesses. That’s where my team is thinking of how we need to bring education first and foremost. We’re a bank and often we go to the solution of financing or investing. But we also do advisory services that play a huge role in education. So, it’s not necessarily first and foremost, the financing that is missing. It’s the education to bring a certain level of knowledge and recognition.”

She reiterated that the industry needs examples of AI adoption and deployment that are not just complex and extensive.

A successful application

At the ALL in AI panel discussion on how AI is transforming the manufacturing sector, entrepreneurs and leaders working in this area spoke about actual implementations and how they fared.

Mathieu Laroche, a senior advisor on technology strategy and transformation, explained that it is important to realize that AI doesn’t solve all the problems. It can also be quite expensive at times and there are technologies that are way simpler to implement.

An example of a successful implementation, shared by Laroche, was a project he worked on at Kruger Products’ Sherbrooke mill, a pulp and paper facility.

The mill, which was built about four years ago, was a great place to start, explained Laroche, as some assets were already digitalized. AI implemented at the mill targeted supply chain demand forecasting. It was used to improve productivity, to ensure that all assets were performing at their highest levels. The mill also added predictive maintenance. AI was used for production planning to plan the production lines for the different shifts.

“The program of course is continuing right now. The business units and team that was built with that program from zero persons to now 15 software developers and full stack data scientists also is now autonomous and scaling that program to other mills,” said Laroche.

He added that being able to maintain a solution and scaling it to other mills is a major achievement.

The human factor

There has been a long-standing fear in the industry that automation technologies can possibly take away jobs. However, industry experts have reiterated quite often that human beings are crucial to the success of technology adoption.

Panellist Sean Clare, co-founder of Pacefactory, noted that operators on the plant floors are often treated as disposable assets, especially when there is a case for cost reduction.

“One of the things we’ve discovered through this use case that we did on understanding standard work, which is the work process that people are supposed to do, [is that] a lot of people, when they start in manufacturing, they feel like they need to be keeping themselves busy. So, they’re working really hard. They’re doing things but they’re not necessarily working smart. They’re doing things and if they’re not doing it in the right order, they’re upsetting the whole process, which has automation, robotics and a lot of other moving parts that aren’t human. So, what we’ve learned is that once we’ve implemented and understand the current state that if people work in the proper process and pace, they actually find they have to work less. The job is actually easier to do because they are working at a slower pace, but they’re doing the right things versus making themselves look busy,” explained Clare.

He added that it is important to equip people with the right tools to ensure that they are successful.

The decision-makers in manufacturing businesses are not usually the people operating the machines and using the advanced technologies. The actual users, such as operators and plant floor personnel, should not only be comfortable using the technologies but also believe in the technologies they use.

Josef Zakowicz, vice-president of corporate development at Canvass AI, shared that in his team’s experience, most plant floor personnel don’t trust AI. He added that it is important to remember that when things go wrong in a manufacturing environment, it can turn into a potentially dangerous situation.

“So, this business of building confidence and trust in what you’re doing is very, very important. As a result, when we work with our customers, we have them focus on the small, low-hanging fruit problems that can be solved. And that builds confidence and it also delivers immediate value. You have to take those small steps before you start scaling anything across an organization. And if you do that well in one plant, (and this is our experience where we work with multinationals) it’s just that one plant gets on fire and it spreads to other plants almost naturally. While the sales process and the deployment process on the first implementation are excruciatingly long, that cycle shortens as you build confidence within the workforce,” he explained.

Not all manufacturing facilities are built equal. Many still have legacy systems and equipment. They have people with decades of experience managing the processes but may not be trained in the most current technologies.

Clare noted that in a discrete manufacturing environment, it is necessary to go through the transformation of using more digital data to support the people on the line.

He added that there is a possibility in the future to extend the whole idea of following standards for work both up and down the value stream by bringing more data sources in. Clare explained that the data could help in learning how to use technology to learn what people need to know at the time they need to know it so that they can apply it.