🎧 Subscribe on your favourite platform iTunes | Spotify | TuneIn | YouTube

In today's episode of the HR Leaders Podcast, we are joined by Sam Zheng, CEO and Co-Founder at DeepHow.

Sam shares his journey from studying engineering psychology to leading DeepHow, a company focused on leveraging AI to bridge skills gaps among frontline workers. He discusses the importance of capturing and transferring knowledge using AI-powered video and the impact of digitizing expertise to enhance operational excellence.

🎓 In this episode, Sam discusses:

1. How AI is transforming knowledge capture and transfer in the workplace

2. The challenges and benefits of using AI to capture and transfer knowledge

3. Practical steps for reducing time to proficiency and improving workforce readiness

Create Digital Experiences With AI-Enabled Recruiting Automation

For organizations undergoing HR transformations, legacy recruiting automation does not support the marketing-based practices of modern talent acquisition or facilitate agility in a context where needs change rapidly. As such, many HR leaders are looking to implement recruiting automation that goes beyond cost reduction and supports agility and innovation.

Download this e-book to learn more about:

● How automation can enhance digital recruiting service delivery, improving hiring outcomes.

● The role of configurable workflows in ensuring adaptability and compliance.

● How portal apps drive efficiency by providing a user experience tailored to the needs of each stakeholder.

● The critical capabilities a modern career site should support to maximize conversion, including content management automation.

● How AI can boost recruiting automation through predictions.

● Key considerations to keep in mind when selecting a recruiting system vendor.

Sam Zheng 0:00

To clear your mission, of course also clear position our mission is to we want to empower your skilled workers to digitize, capture, share and transfer their knowledge and to help them excel to achieve operational excellence. And what we're uniquely your position is particularly your knowledge of capturing organization at transfer or L XP. IMS some other vendors in this space where we talk again connect back to my background from psychology is something we think very important to capture experts know how through demonstration.

Chris Rainey 0:40

Hi, everyone, welcome back to the HR leaders podcast. On today's episode, I'm joined by Sam Zang, who's a CEO and co founder at Deep House. During episode Sam shares how to harness the power of AI to bridge skills gaps amongst frontline workers, how to capture, manage and transfer knowledge using AI powered video, and how to transcribe and translate expertise to turn hours and hours of work into minutes. As always, before we jump to the video, make sure hit the subscribe button, turn on notification bell and follow on your favorite podcast platform. That being said, let's jump in. Hey, Sam, welcome to the show. How are you? Good, how are you? I'm good. How you been? We've been up to crack crack

Sam Zheng 1:26

fantastic. Being being busy. As you know. AI is advancing You're fast. And we you know we want to connect the technology apply the technology to help people and particularly your to help scale workers.

Chris Rainey 1:42

Well, on that point, you've been on an incredible journey. And so I'd love to give everyone a bit of a background. So tell everyone a little bit more about your personal journey and how this led to the work that you're doing now deep how sure

Sam Zheng 1:54

you are I actually can I study engineer psychology, and always interested in understanding people helping people, but seeing how technology is really transforming every aspect of life and work. And that's why I was I was thinking, hey, I want to study the intersection between technology and people. And then my PhD many years ago from University of Illinois, at Urbana Champaign. And after that I joined Siemens corporate research. And that's an innovation center for Siemens based in Princeton, New Jersey, and spent over a decade there and leading and driving many digitalization project, and many now, so called industrial 4.0. projects. And again, a lot of focus is trying to bring in more automation and robots to the shop floor to the manufacturing to the industrial environment. But what we will quickly realize is we're in talking to folks on the shop floor and talking to folks in the field is that Yep, we are definitely need more automation. But we don't have enough skilled workers. Right? A lot of experts are leaving, take away their knowledge. Young people learn differently if we were struggling and trying to write to to train people. And that really made us think, Hey, we got to do something different. How can we leverage technology to help people to solve this so called Skills Gap, your topic?

Chris Rainey 3:31

So really, you kind of identified that challenge? And that was what inspired you to start the business? Absolutely,

Sam Zheng 3:39

absolutely. That's where we, that inspire us to build the power again, we see multiple trends going on. Right number one technology is advancing so fast, right today, AI, of course, your industrial 4.0 digitalization, that means they will need to learn more and need to be with skilled and upskill. And so in this your, your skill technical domain, and meanwhile, right millions of experts, they're approaching retirement but the baby boomer generation, and they spent years on the shop floor and building up their experience, their knowledge had not been effectively capture and transfer to the younger generation of workers. And speaking of the younger generation workers, the learning paradigms, something has shipped, right? They are no longer interested in reading, instruction manuals, SOP documents, and all these written format and they'll be scrapped. If they want to learn anything they prefer go online. They'll do you to go to tick tock they want to see something visual video, more interesting your counter. That's where we see that we opportunity here. Can we elaborate the right the format that the new way of learning applied in this environment, and that will re inspire us to create the power and of course, the whole content creation, video creation, it takes a lot of time, and a lot of your skill require and our ideas eight, can we leverage AI to make this process a lot easier, right can can transcribe your experts, you're speaking in different languages in very noisy environment, and understand what they're doing. And also, you're organized all these knowledge and turn complex the workflow into these step by step instructions and make it very easy. And try to translate that into different languages and make it very easy for workers to learn.

Chris Rainey 5:46

How do you describe where do you see deep How fitting into this ecosystem for HR leaders? Do you do you do to describe yourself as a LSP? Do you describe yourself? You know what, what what? What I actually know, firstly, before we get into that, tell everyone obviously, a little bit more detail about deep how the solution, you know, what makes it different? But also, yeah, more specifically, the customers that you're serving? Because I think that's important. Yeah,

Sam Zheng 6:15

totally, totally, we have a clear your mission, and of course, also clear position. So our mission is to, we want to empower your skilled workers to digitize, capture, share, and transfer their knowledge and to help them excel to achieve operational excellence. So and what we're uniquely your position is particularly your knowledge, the capturing, organization and transfer, we actually work with our L, XP, IMS, and some other your vendors in this space. And we'll we'll talk again, connect back to my background, and from psychology is something we think very important to capture experts know how, through demonstration. Because today, if we look at how knowledge is captured, typically, in this space, they're captured in written and format, sometimes with pictures and so called SOP standard operational procedure. And you either have a technical writer, and they need to work with expert, they actually go talk to the expert, and then an expert have a need to based on their memory, to share, hey, this is how I'm doing things. But there are a lot of details are often left out. Right. And because these are skill type of your work, procedure work. And in fact, it's a lot easier to demonstrate how you're doing things, there are a lot of rich information, nuances in the context. And that's the power we want to capture. And we'll capture really like well, I cameras, and it will just open this up. Either they can do self recording, or someone else can can can record what they're doing. So we want to turn this rich demonstration and labor into the instruction. And by the way, and that's why people also enjoy watching. Right when you're watching video, because it's just a lot richer information. It's only so

Chris Rainey 8:33

true. Yeah, it's so true. Because, you know, I, I read a I read a when I first built, I'm just gonna give you a silly example. But it's just I think will resonate. Like when I first built my first PC. Right? I read the manual of how to build a PC, which is very difficult, right? Where all of these components go but what I did actually is just watch someone on YouTube building a PC and I built along with them. As we were putting in the CPU I put in the CPU, as they were putting in the RAM I put it in the RAM as as the when I put it into the PSU to power supply I put into power supply. Absolutely right and that's and it's it's similar with my friend works for Ford automotive, and he builds engines there and he learned to build the engines while watching videos and on all just watching other people in person. Right? Yeah, but so how you mentioned videos, talk more about how you're doing that and also how you then use leveraging AI on top of

Sam Zheng 9:44

like how you said it, you can reflect your own experience, in fact that you see that apply to all of us. And this actually there's a plenty of research, real papers and findings to back this up. All right, and particularly to learn your skill type of work, and learning from demonstration. And this is so cool, imitation learning, right? You can learn much better by watching other people doing things. And that also knowing that you will learn quicker, and then we pay knowledge, you're faster, longer, but also people will develop better confidence,

Chris Rainey 10:26

you're also applying the knowledge, you know, you're, you're not just consuming it you're applying. And that's why that's the very, yeah, yeah,

Sam Zheng 10:35

that's why in fact, I have seen it even at home, I've seen the experience, my wife is very organized person, when we buy some home appliance or other digital device, is she typically when you're organized all the manuals and put into a huge binder. The same way these get broken, we never flipped those manuals, we go to YouTube, you're looking at someone you have done something your cylinder. And that's why you even as soon as we saw some of our technicians, engineers, instead of going through, so go into the field, they actually will receive, I mean, they will bring with them today is no binders, they digital copy, but we say is a digital replica of your paper values, right? So PDF they they load is 100 of these manual in the thumb drive and go into the view. But yeah, they often they will go go online, try to look for something or they have to call HELP. So that's why we see you we important opportunity that the key thing lives at then how can I take that experience to the industrial environment to the manufacturing environment, we spend years of work and first we have to work closely with skill workers in the field, understand the manufacturing environment, for instance, right? This is not like if you want to recall this video, and this is now in a nice, you still do the environment like currently we are in? It's very noisy. There are a lot of things going on. And people you're speaking it with strong accents. And so that's always the first thing and how can we develop an application and mobile app make it so easy for people to use in that kind of your environment? And then we saw the noise issue, because sometimes the environment could be we will have people need to wear all these ear protection. Right and cargoes. So so the first thing is right, so we will we custom, our you know, solution for those environment. And then the second part, these are not, I mean, they are skill experts and their skill your workers, but they are not expert in video editing. So how can we use AI to make this process so easy for them? So they're not afraid? Because there are a lot of those video editing software is full of gauges and flow features. It's integrated, intimating to them. So that's why

Chris Rainey 13:28

it's so easy is that where is that where you talk about the auto segmentation? Is that correct? That's why

Sam Zheng 13:33

That's why we built our AI. In fact, when people start we started five years ago, we're already leveraging AI and use AI to do your transcription. And that's the first part and in very noisy environment I mentioned but more importantly, is use AI. Wearable. Wearable, early on, we already started with multimodal your understanding. So we you look at the your AI and if it has an ankle Stephanie so she will understand the experts you're talking from audio and transcription, but more importantly also looking at the you know the action and sequences and that's when the computer vision and combine all these information, turn this complex workflow into key steps. So that we call absolutely auto workflow segmentation and again turns

Chris Rainey 14:32

super helpful because you see the same thing again on YouTube. He does the auto segmentation on videos on YouTube, we

Sam Zheng 14:41

did that even before YouTube. If you look back, you will see actually YouTube did not release that feature. At the beginning YouTube still need people to go there and specify the timestamp

Chris Rainey 14:59

My team did that for many years. Yeah, that's right. It's a lot of work. It's a lot of work. And I think one of the main things that you've already mentioned, though, is because of the global companies you serve. You know, companies like Unilever, right? And it was a client of yours. Yeah, that's one of our customers is very important for them to have the auto translation. So they definitely have a global workforce. And you can automatically translate those videos into any language. Is that just text? Are you working on audio as well? Are you working on voice?

Sam Zheng 15:31

We started with pecs, and you're absolutely right. In fact, you know, to make the knowledge not only capture digitize thing and make it accessible to the global workforce, and that's so critical. And we we work with, and today we work with over 100 an enterprise and SMB customer, and many of them are really global company, and I you mentioned you deliver and also Anheuser Busch InBev, by Stanley Black Decker, they may literally have workers in every continent, speaking different languages. And that's really the key thing. So first, we start with the, your translation packs, and why all because again, in a lot of these environment is very noisy. So people actually cannot hear, like I said, they need to hear, they need to have the hearing protection. So So that's why, again, translation into text is critical. That's what we're working on. But of course, we also can do the audio translation as well. Oh,

Chris Rainey 16:39

you can already? Yeah, yeah. Really? Wow. That's incredible. Yeah, I'd love to experience that. Because I think that's the part that is one thing to do transcriptions into different languages, but the audio transcription is incredible. Is that like using a user? Is it your own proprietary or using chat? dBc for to do that? Like, what? How are you leveraging AI for that? Yeah,

Sam Zheng 17:02

good question. It's a combination of both. So like you mentioned, your tech GPT, particular GPT. Four. So these are large language models powered by your transformer are these, the transformer is a deep neural network. And that was demanded in 2017. At that time, we actually started the company in 2018, we're already leveraging the latest technology. You're absolutely right. So so these IGBTs are transformers, because they are inspired by human brain, human neural network. And by the way, we, our brain can do many things, even though there's only one, your neural architecture, right so we can hear understand things we can see and also make sense of the visual information. And then we can connect multi modality excetera. So that's what today the technology can do. And same thing for our technology. That's why we started with multimodal first. So a lot of those critical, the critical your algorithms were developed by ourselves, like for instance, auto segmentation, but we also take advantage of a lot of the state of the art in particular today, like for instance, voice generation and translation, these are already sort of probing, and then we tap into your state of VR language, their API's, and we don't need to build everything from from from scratch. And but one key thing is because we're working with, again, industrial customer. So and privacy is very important, a lot of enterprise, your topic. So that's why we put a lot of your emphasis into our solution, we make it secure, right? And that's privacy, your top is making sock two, type two, well sock to type two, your compliance at your very earlier, because we want to make sure right, our customer than then is comfortable, our users, they're comfortable, their data is not just going to

Chris Rainey 19:10

open up I can't You can't get around that. Talk to me about how this is helping companies reduce their time to proficiency. And yeah,

Sam Zheng 19:20

absolutely we easy. We have lots of examples where reported for from our customer and user, but we also conduct a systematic study. So very early on, so we tried to your compare. So there are multiple parts and the benefit. Number one is how we can actually reduce experts time and make them more efficient. And we see that because today expert at least when a manufactories we visit is span at more than 20% of their time, and repeatedly you're teaching new employee some even mundane tasks. And only this is not the good use of the time. And also a lot of new people that are afraid to ask questions. And now with digitize that knowledge and that can make the interaction with the expert a lot more efficient. This young workers, they can watch your video, they can speed this up big

Chris Rainey 20:19

Expo slow down to expert, you now can spend time on more high value tasks, absolutely

Sam Zheng 20:24

higher level more complex topic. And we see that that actually clear the saving experts time and improving also new employees onboarding. So across this, we have seen your 30% 40% and quicker onboarding. And for for for tasks. Because again, a lot of

Chris Rainey 20:46

you can scale it right, you can now do at scale, which you can't do one to one one to one once you can scale.

Sam Zheng 20:54

We all know right is that once information is digitized and organized and all more importantly, easily accessible, then, then then then that everyone can just use iPad, iPhone and whatever devices they are allowed to use on the shop floor at home, they can access this. And we're

Chris Rainey 21:13

going to go into that a little bit more sorry to cut you off. Like one of the big challenges. And this is why I was excited when we first connected is I told you we I spoke to many companies to have a real challenge connecting with their frontline workers. What is the experience for the frontline workers? Do they have an app on their phone? Like what does that look like? How are you connecting with those frontline workers?

Sam Zheng 21:38

Totally. Yeah, that's actually that that's really the power, we have years of experience working with your frontline workers and workers on the shop floor environment. So you're absolutely now we seeing it's becoming now a lot more common compared to a couple years ago and their iPads and on the shop floor. And then technicians they already have your mobile phones, your Android, all these devices is definitely just becoming a lot more popular. But one key thing I do want to mention when we started at a time again, people all have fear about AI. But the fear is Oh, AI is coming like self driving car is coming. Robots are coming. They're going to your workplace, we will take all the jobs and started with motor skill workforce. Right when I first but it turned out you have all these years I the woman right? You mentioned about LGBT. And it's actually threatening more on the white collar, your worker jobs, and all these skilled worker but at the beginning, that's why we we will work closely with our customer with our end user, we may not understand hey, this is a powerful technology. And although this will potentially replace some of the tasks, but you can also take advantage of this to help you get better. Right. So technology always has two sides. And today again, now many of them they that's why they love working with us. They love our AI they are as this thing called Stephanie they love Stephanie because definitely is here to help.

Chris Rainey 23:24

Can they are Stephanie questions to instantly find content? So yeah, absolutely. So rather than searching for a library of of content and resources, can they say Hey, Stephanie, show me training on x? And it's definitely finds them that.

Sam Zheng 23:40

So, uh, yes. And we're also working on the latest Gen air technology to take Stephanie to the next level. So the first thing is, Stephanie can already do very comprehensive, your search. And, of course, people you're checked up all these transformer just like Google, it can pretty much it can only do the keyword

Chris Rainey 24:07

or, like smart. So

Sam Zheng 24:10

yeah, all the UC related to your documents. But today, yes, we're already releasing beta. It's a platform. Basically people just ask questions, and then we'll return answer. But we're we want to make sure because these are the environment we know AI can hallucinate, right? And these are the environment we only make sure the data is accurate and reliable. And that's where we leverage this retrieval, retrieval augmented generation wreck. And we always want to make sure hey, I don't make up things. Look at your the all the data we already have digitized and look at the expert. Always make sure you do the citation just like I would actually teach my students because I also do When part time you're teaching advising students, you cannot just write something, making the claim without backup with Africans, right. So that things. So yes, that definitely this is something we want to make knowledge very easily accessible. Sam,

Chris Rainey 25:14

I know I left, I know I have to let you go, because we're coming to time but you know, what would be your parting message to the HR leaders that are listening? And where can they connect with you in the team, if they want to reach out, totally

Sam Zheng 25:27

go to our website, deep how to account like the t h o w n, where you can find all the information and then the counter. So we came up with HR leaders and implicate in the skilled workforce domain, we actually see that something is emerging. So the HR leader will team up with the operational leaders, and create even a new organizational unit called workforce readiness, because HR leaders definitely has all the expertise, organize all the training material and understanding different job title excetera recruiting and people and but operational leaders, they are the ones working with machines and dealing with all these assembly lines, etcetera, and make sure things can produce they have the domain, you will know how so? So the power actually we connect HR leaders and operations leaders and they work together with the frontline workers. So yeah, we're always excited and looking forward to the opportunity to to discussing all the HR leaders.

Chris Rainey 26:42

Listen, I love what you're doing. I think it's incredible. And anyone who's listening right now it has a large population of workers, you need to be speaking with with with with Sam, and the team is a lot of what you're solving right now is kind of the challenges I'm hearing day to day from especially as I mentioned those companies with a big population of frontline workers and that as someone who kind of pays attention to the market there, there are not many solutions out there. And that's why I was really excited when we first connected. And I'm genuinely mean that as well. So anyone who's watching, we'll put a link below to Deep House website, also a link to Sam's LinkedIn. If you want to connect with him directly on LinkedIn. You can check out the clients they're working with, as you know, Stanley Black and Decker Unilever, the list goes on. You're doing amazing work. So I'm really happy to see to see you doing well and I look forward to catching up again soon. Thanks so much.

Sam Zheng 27:40

Thank you so much, Chris. Well, we appreciate your time.

Chris Rainey 27:43

Thanks

More from the HR Leaders Podcast