Podcast: AI-assisted, cloud-connected radios are reducing language barriers on the factory floor

For us, step one is capture this dataset that is untapped, essentially, and the way that it works is when you replace a traditional radio with our cloud-connected radios, every spoken word is data, right? It runs through our cloud, it gets transcribed, it can be... translated, which is one benefit to the workers, is now language barriers don't have to be a barrier anymore. With our solution, for example, if we had five different workers speaking five different languages, they could all be speaking in their native, natural, most comfortable tongue, and our radios will, through the Cloud, leveraging AI models. instantly translate those in real time and play them out in the right language on the other side, right? So that's an immediate benefit that turning voice into data can provide. when married with AI. 

But then, if you think about, again, the work happens verbally throughout a day, well, at the end of a lot of frontline jobs, you have to write something like a shift report, right? Kind of like, what happened on my shift? Did this machine have a maintenance issue? Do we need more supply of X, Y, or Z? That is a manual and frankly, painful task at the end of a long shift when people are tired. They'll either not do it at all because they're tired or they'll do it overly quickly just to get it done. That data is unclean to your question. But what we do is we capture the data passively. It's just happening throughout your shift as you're talking. We leverage AI to then synthesize, essentially write a shift report for you. and then you can audit it to make sure that our capturing is clean in that sense, and then it gets uploaded to whatever system of record, ERP system that the plant may use. Those are some examples of how we're capturing data that's never been captured, and then cleaning it in a way that I think ultimately saves the human time. and hopefully saving them a lot of pain in that example as well. So those are just a couple of examples of how we're approaching this problem.

SA: Very descriptive. Thank you. And looking at my second question here, it might be a little repetitive with what you just went over, but I wanted to point out that in your piece for us, you were pretty descriptive, but loosely descriptive of the types of AI-assisted technologies you were talking about that could help frontline deskless workers. Can it be a more specific in description with us here, whether you're talking about relays, products or offerings or others? What are the current generation of AI enabled communication tools, Chris?

CC: Yeah, I think we're like pretty early on the evolution of AI enabled communication tools, so this is pretty, you know, white space, if you will, that we're starting to fill out. But, you know, and it really is not the only such solution. But, you know, just to briefly recap, you know, of course, I mentioned language translation is one benefit. And I think even as consumers, we're starting to see more and more announcements. Apple just announced some stuff with their earbuds and T-Mobile as well. So I think, you know, the problem of language barriers in various contexts, consumer and frontline is being attacked through leveraging AI models. We're at maybe the forefront of that for front-line workers. That's one example. I mentioned, again, as voice becomes data, you can start to capture what's happening in a shift or in a workflow and do things like automate shift reports. 

But another example that we haven't touched on yet is If you've ever used radios and particularly as a manager, there's this notion of channel scanning, which means historically that you are flipping the channels manually of a radio, listening for something important, essentially doing a human manual keyword spot across all these different channels. obviously, you can probably envision, if you've never tried that before, that it's fraught with a lot of randomness, right? Am I on the right channel? And therefore, you might miss the key signal that you're looking for. Something urgent is happening, something needs a repair, someone's in distress. 

And so we think that's a perfectly ripe use case for AI to help, because why rely on a human manually scanning channels? By the way, this is a common use case for security teams if they have a security center, right, where they're monitoring all the radio channels for security issues. But any number of situations where channel scanning is applied, we think a better solution, be it from relay or otherwise, is for AI to scan all the channels and listen to every word, and so you never have to miss that important message that you otherwise might have in the old way. Then be notified through the AI of that important thing that you're looking for. We think that's a really powerful innovation that AI can bring to a number of different kind of jobs and use cases. Ultimately, that not only will help productivity, but safety. Again, if someone's in distress, never miss that distress call, if you will, because you have AI watching over and scanning all those different channels for you.

SA: That's really interesting, Chris. One other point I wanted to ask about is, I wanted you to elaborate on this statement from your February 9th column. The conversation must move beyond automation alone and start focusing on how AI can augment the way frontline workers communicate and make decisions. I think you're talking about exactly that in your prior answers to the questions, but can you elaborate a little bit more?

CC: Yeah, so AI can really, I think, augment, you know, again, we've talked a lot about the communications, but also decision-making in that the human signal that we've been referring to on the frontlines is not only voice as we're referring to, but movement. That's the other ground truth data that we like to refer to. What's happening in a frontline environment is what people are saying, but also what they're doing and where they're going. Solutions like Relay also track location on not just a GPS, but an indoor basis, and so we can know what part of the factory you're in, have you been in an environment too long? How can that be useful? Safety, for example. In a lot of manufacturing environments, there's lone worker situations where there is a danger in a particular place, When we know through location that you are in this dangerous place, and we also know through other means that you're alone because there's no other relay device near you, well, now AI can help monitor that lone worker, call it automated check-ins where we can essentially, like a manager would, make sure that that person is safe and secure. And then if that person doesn't respond, then basically the AI can then notify the actual human in the loop and help marshal a response to that situation. So that is a safety decision that I'm painting here, but you can imagine many other similar use cases that can be solved if you have the data from both voice and location.

About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.

Listen to another episode and subscribe on your favorite podcast app