What happened to the promise of Artificial Intelligence (AI) in automotive manufacturing? When did the mention of Artificial Intelligence become an eye-roller to the manufacturing C-Suite? When did it become a quixotic undertaking? It’s almost as if ‘Artificial Intelligence’ has become a couple of dirty words.

What was the promise, exactly?

I am going to mix concepts of Deep Learning, Machine Learning and AI together, because either you know the difference or they are closely related enough for you. The biggest problem with AI is that manufacturers have not stepped back to understand the art of what’s possible, and many industry consultants have not been able to create solutions. And this, the lack of solutions, is the reason why they have become dirty words – the industry experts have not yet brought solutions that fix problems en masse.

Promising Solutions or Industry 4.0 Hype?

As I think back to 2015 and 2016, I remember that predictive maintenance was one of the hot areas associated with AI and Machine Learning. Here was the premise – collect all of the data surrounding a given machine type and usage across a network, plot out that data, and use it to tell you intelligently when the machine will actually need maintenance, rather than relying on the owner’s manual or your plant employee’s best guess. Then, once the machine data reflected that maintenance was due, there could be an automatic notification created and sent to the vendor to come perform the required maintenance.

Do you use this in your manufacturing company?

I’m not asking if you’ve heard of it. I’m not asking if you did a pilot program on a couple of stamping machines.

Do you use AI in your manufacturing company?

Have your worst fears come true? Is everybody out there using it and you are the one that is left behind? Are there big social events where everyone is celebrating massive cost savings from their predictive maintenance solutions over wine and some kind of finger food, and you are on the outside looking in?

No, not really. Yes, these initiatives are growing – slowly – but not at the rate that was expected. Industry 4.0 hype, in general, grew massively, far outstripping the actuality of what was being done. Listen to the wording of this quote from Business News Daily. It’s related to augmented reality more than AI, but the same concept applies. I won’t tell you the date of the article just yet, but carefully consider the wording.

“Mixed reality is also a major component of Industry 4.0. Big companies are already issuing mixed reality devices like helmets and glasses to employees in hopes that the increased communication and visualization of contextualized data will boost productivity and intelligent decision-making.”

Wow, big companies are already issuing these devices? So, if I walk around in the typical manufacturing company, I am going to see a bunch of specialized glasses and helmets on plant employees?

That article was from November of 2017, folks. So then everyone must be using this stuff now, right? Nope. There are pockets of usage of things like augmented reality and machine learning, but they are pockets. There are pockets of areas where predictive maintenance is used, but, as one of my colleagues, Glenn Graney, QAD Director of Industrial and High Tech, advised me, companies that he talks to often think a disciplined Plant Maintenance program can be more effective and easier to achieve than adopting the complexity of Predictive Maintenance toolsets.

So, what do we do now?

Is there Still a Promise of Benefit from Artificial Intelligence?

I believe there absolutely is. Think about the most successful areas of AI. I will start with autonomous driving testing. Cars collect data over thousands of hours of driving and millions of miles, using machine learning and deep learning to get better and better at decision making. It is already yielding millions if not billions in revenue and will jump to trillions.

It required heavy investment, lots of patience, and some incredibly smart people that understood the data elements, variability, and the technology platform they were developing to master it. Autonomous driving is now far safer than manual driving per mile driven. It is our inability to trust that has so far precluded far greater adoption of the tech. In my opinion, herein lies the core difference in why we haven’t seen an explosion of adoption in manufacturing companies.

No one seems to have investment dollars, patience, or the right skill sets in their manufacturing departments, along with a sage-like understanding of the applications and data to really drive adoption and value in manufacturing. And we see existing companies already starting out with near insurmountable challenges just in core fundamental items, let alone these advanced concepts. For example, most companies don’t have a single type of Bill of Material (BOM) construct. They don’t share a commonly governed set of master data – item master, vendor, customer, chart of accounts, etc. They have multiple code sets and versions of ERP and MES software, and different PLCs and sensors capturing data, so that if they ever did get patience and investment capability, they would be unable to build and maintain all of the cross references and algorithms required because of all of the different systems and master data.

Looking Forward

Startup companies will have a massive advantage over entrenched companies, as they are often making the right decisions from the outset. They know the power and necessity of convergence and simplicity – they are selecting just a few cloud-based systems, establishing master data governance and best practices from the beginning. They are building data lakes that then sit upon those governed applications, and they will drive benefit.

Artificial Intelligence, machine learning, and deep learning will have explosive growth, but companies need to take the steps to build the foundation necessary. Highly governed master data, best practice processes, and simplified IT architecture are required. And companies must hire and develop, or internally develop employees that can understand the data to a table and field level and full business usage level, along with the necessary IT investment. And it will take money, patience, and force of will to make it happen.

Discover more Dirty Words in Manufacturing by listening and subscribing to QAD Podcasts.