How Machine Learning, Wearables and Smart Devices Could Close the Skills Gap Once and For All

This technology could disrupt industrial work the way GPS disrupted asking for directions.

Image courtesy of Contextere

Image courtesy of Contextere

Machine learning, data collection and analytics are becoming commonplace in the management offices of manufacturing and industrial enterprises. These tools help drive improvement and inform decision making. But what about the little guy?

On the shop floor and on the field, workers are less likely to have systems in place that use advanced technology to support their knowledge and skills. However, digital tools can provide the guidance and information they need to complete new and unfamiliar tasks faster and more reliably.

Carl Byers, co-founder and Chief Strategy Officer, Contextere

Carl Byers, co-founder and Chief Strategy Officer, Contextere

Contextere is using machine learning, wearables and smart devices like smartphones and augmented reality glasses to bring needed information to the blue-collar worker in the field. For example, a  technician performing maintenance on a machine in a factory can access the machine’s maintenance history and sensor data, and instructions for performing the needed maintenance procedures via their smartphone, on-site. 

Engineering.com sat down with Carl Byers, co-founder and Chief Strategy Officer at Contextere to find out what they’re doing and how it can impact the manufacturing and industrial workforce.

The skills gap is a significant issue in manufacturing right now. To start us off, how would you define that problem? In a nutshell, what is the problem of the skills gap?

Well, there are a few factors contributing to it. One is with the changing workforce, generally. Retirements are accelerating, and new entrants are coming into the workforce with a long gap in knowledge. As I said, it’s starting to pervade the workforce much more than it has previously. That gap in knowledge is rising. I think that’s the first issue.

The second issue is, if you look at the required skills for the workforce, those required skills are changing very quickly, as well. With the combination of new tools and technology, then automation and all these kinds of things in the workforce, the skills that people need to be functional the minute that they arrive are much more complex, and the more senior people have less native skill in technology.

So, you end up with a lack of people that have the ability to hit the ground running, just because there aren’t enough of them coming out of schools and training. Then, once they’re in the workforce, they don’t have the support mechanisms from the vast body of knowledge of more senior mentors that they may have had in previous years and decades, as well. They don’t have that structure to be able to learn on the job, even, as fast as possible. That results in this gap in performance and gap in productivity that organizations are trying to fill both by getting people into the workforce that are more effective, as well as finding new ways for them to actually be more functional on the job more quickly.

And that’s where you come in.

That’s correct, yeah.

Contextere has this concept of blue-collar AI, machine learning based curated guidance on mobile and wearable devices. Are we talking about things like augmented reality glasses and smartphones? Can you tell me a little bit about what is blue-collar AI?

We’re focused on the hands-on skilled worker. The field worker that’s putting their, as we like to call it, warm hands on cold steel, going out to fix and maintain, install and operate complex equipment in the field. That tends to be characterized as a blue-collar workforce rather than a white-collar workforce.

The skilled technicians that are going out and doing those jobs can and should benefit from the advancements in artificial intelligence and mobile technology much more than they have and are, at this point. Their jobs and the tools that they’re using haven’t changed, fundamentally, in many, many years. They’re being passed by, in a lot of places, as the headquarters and the white-collar environment incorporates more AI and analytics to help them make better decisions and operational decisions. That same value of making better decisions, having information at your fingertips, accessing the vast body of knowledge that might be available across an enterprise, that same value has the potential to impact a blue-collar field workforce. It’s not currently being delivered out to them in an effective way.

Our premise is that if we can understand who somebody is and where they are in the field, then we can understand what information is important to them, and what guidance should be delivered to them on their smartwatch, on their smartphone or on their wearable device to be able to be more effective and more productive in that moment in time. Rather than taking traditional mobile connectivity solutions that people have often provided to those blue-collar workforces in the field, like, “Here’s a laptop,” or, “Here’s a tablet. All of your corporate information is on it. You go figure it out now,” what we’re saying is there’s a smarter way to do that.

We should be able to understand who you are and where you are, and then, from that vast pool of technical information, deliver you the stuff that you need, in the moment that you need it. That’s where blue-collar AI comes in.

In the conversation around the skills gap, a lot of solutions have been proposed. Things like bringing in more automation; or education and upskilling; or things like STEM awareness for students. Would you say that these solutions are viable?

I would take the approach that another piece is needed. I think we have to do a lot of things, and a lot of things will happen just naturally. Automation will continue, machines will get smarter, ways to automate efficiency on equipment will improve. Connectivity across equipment, and then self-optimization around that will continue to improve. We know that’s going to happen.

Industries, to differing extents, will adopt those technologies. Depending on whether you’re a small or medium manufacturing enterprise, or if you’re a Foxconn, a massive manufacturing enterprise, you’ll have different abilities to adapt to and adopt those technologies. Our machines will get more efficient and more effective.

What we’ve seen continuously is that no matter how smart those machines get there’s a still a need for somebody to go and fix them, maintain them, or install them in some way. Those things don’t happen by themselves. Maybe if you can predict when maintenance is going to happen, because you have more sensors on equipment collecting data, but you still have maintenance actions or repair actions that have to happen because equipment just fails.

The level of knowledge that you need, as a technician, has to be even more advanced because you’re not going to see the same kind of daily failures or daily maintenance tasks that you might have seen historically, where you could just do all of this stuff by rote memory. That requires you to have more access to information that might be more unique, related to the equipment, related to its operations, related to the broader knowledge across an organization.

On the issue of STEM training and skills training, I agree. Digital literacy, generally, needs to expand. We need to continue to adapt and accelerate the level of digital literacy in our school systems, from lower grades right up to postsecondary. There’s no question about that, but it’s not a panacea. It’s part of a whole package of learning and evolution that people need. Those blue-collar workers that often used to go into blue-collar workforce, because they like to fix things and get their hands dirty, and get their wrenches to fix the brakes on their cars on the weekend… What they’re seeing now is that that’s still the case, but the digital component of what they’re working with is much more advanced, and they can’t get away from that. They need to build a deeper knowledge in that. We know that that’s the case.

The challenge is with the existing school structures and the postsecondary vocational institutions. As they continue to advance and evolve, they’re still not creating the volume of skilled workers that are forecast to be required by the market. One study that was released last year showed that in the U.S. alone, 2.4 million manufacturing jobs are going to go unfilled in the next decade. 60% of new jobs that are going to be created are going to be unfilled, and that’s largely because the pipeline of education is not going to be able to push the people as fast as possible and give them the skills that are needed.

Again, you could say, “Let’s just improve the educational system.” But, again, by itself, that’s only part of the solution. What we believe is there’s great ways to combine various things. There’s going to be automation, there’s improvements in educational systems. But, at the same time, what we could use is technology to be able to deliver on-the-job training and on-the-job learning much more quickly and much more effectively.

Let’s imagine way back to the old model, when there used to be apprentices and master craftsmen for every job, and guilds helped people get into the workforce. What we see is the digital equivalent of that. With something like Contextere and our Intelligent Personal Agent, it’s a digital master craftsman that can guide an apprentice or a novice on-the-job to accelerate their learning and accelerate their competencies. When they get they get out of school, they’ve got a certain amount of skills. Then, on-the-job, they can learn that much more effectively with the support of the master craftsman.

Would this technology support be used at the beginning of my job, say, as a maintenance technician and I would eventually ‘outgrow’ it? Or, would I use it throughout my time in that job role?

We see this as a lifelong competency development capability, when you look at it in a larger potential. It starts from your first moment on the job. We understand who you are and where you are, so we can deliver you guidance associated with that. What we want to do is tailor that guidance to your competencies. When you’re a novice, your competency is relatively low. Or, if you haven’t seen a problem before, you’re assumed to be a beginner addressing that. After you’ve seen that problem a few times in that environment, the Intelligent Personal Agent delivers guidance based on learning methodologies so that you are, theoretically, absorbing and capturing that knowledge much in the way, again, a master craftsman would guide an apprentice.

Over time, your learning evolves. Your competencies evolve. What we can do then is by monitoring and tracking your activities and performance, we can validate your skills, your competencies, your credits as you grow and create, both for yourself as well as the companies that you work for. This can be a lifelong learning competency and skills mapping capability that you can take with you. As you change jobs, you’ve got a digital verification, so to speak, of your evolving competencies and your validation of those competencies, going forward.

Implemented properly, it has the potential to help individuals upskill, cross-skill, and multi-skill across and throughout their lifetime as they’re in the workforce, and to do it very effectively.

Will this make employers more confident to hire lower-skill candidates?

We believe it will, for two reasons. First, what happens now is if they feel that they need access to more people more quickly, they have to put them into a training program. What we often see is those training programs are very long, and it takes people directly out of the formal workforce, so to speak.

With the turnover that often occurs in industry, as well, what we find is that they’re in a training program longer than the turnover is. The example we have is Schlumberger. It takes them roughly seven years to develop a fully autonomous oil field worker. That same worker turns over and moves to another company within two to three years. So, Schlumberger is investing in developing that worker, and at some point in that investment process, that worker’s going to walk and move to another organization and take their learning, the Schlumberger investment, with them.

We see that regardless of what the training type is, that people are moving to new jobs before the investment that the company’s putting into them is actually reaping value.

So, complementary to that, if they can give them some traditional training, or have them more skilled coming out of school, that’s a great thing. If, they’re actually learning on-the-job and being guided continuously and being validated in that process because the intelligent agent is verifying and validating both the compliance and the competency level as they’re going, in terms of functioning to competency, the employer has a digital compliance record of the performance of the individual that can be validated and increase their comfort level that that person is learning effectively on-the-job.

Carl, what does AI mean today? Because we don’t have true AI, and it’s a bit of a buzzwordy term.

I think you’re bang-on, there. I’ve been in conferences where some very senior people in a large organization have stood up and said, “If you’ve been using Excel, you’ve been using AI all this time.” Well, maybe not quite so much, but we see the term being used all over the place.

What we tend to focus on is a subset of AI, which is machine learning. That’s, basically, using the historical information, finding patterns in the historical information, and matching those patterns to new potential use cases, and then learning and evolving based on that. It’s not these generalized artificial intelligences that they talk about in the science fiction books. It’s highly focused on existing data, and extracting information and knowledge from this existing data, and then extending that to new potential use cases. That’s what we’re focused on.

What devices are being used in the field for this system?

In our case, we currently deliver guidance onto smartwatches, the Samsung Gear S3 in particular. We also deliver the guidance on smartphones from various sizes, and tablets, if people want it. We also deliver guidance on the Microsoft HoloLens.

In each of those cases, those are obviously very different screen sizes and display formats. What we do is we tailor the type of guidance and information that’s available to the user’s environment. Obviously, if you’re on HoloLens, you can actually extend the guidance to have a “see-what-I-see” conferencing capability, and you can view more immersive or semi-immersive animations or videos to be able to remind yourself how to fix something. If you’re on a smartwatch, you can get cues and guidance and physiological data and things like that and smart IoT data from the devices. So, it ranges. What we do is we tailor that information so that it’s consumable, and can make a cognitive impact very quickly in the machine critical environment that we work in.

It sounds like this could be disruptive as GPS was for paper maps or asking for directions.

That’s a good analogy. In the early days of the company, that was actually the one we pulled out as a good comparison. Before GPS, you used to have to carry maps with you. Then, with GPS, that information was at your fingertips. Then, with things like Siri enabled or other intelligent agent enabled capabilities on your smartphone, your reminders would be delivered to you based on your location. You know, the example of “Make sure you pick up milk,” when you get to the grocery store kind of stuff.

Contextere has a similar role in the technical blue-collar workforce. We can track where you are and what your role and skills might be. Then, if we know you’re in front of a piece of equipment at a customer site, then we can make sure that we deliver you the right information about that customer that you need to know at that moment, the historical records of maintenance on that equipment, and any service bulletins and any other data that might be important to you at that moment in time. You’re not searching for it, you’re not spending unproductive time trying to find the information that you need.

For more information about using smart devices and for maintenance and training, read PTC’s Vuforia Brings Augmented Reality to Maintenance and Training.