Episode Categories:
Resources:
Join the discussion: Ask Jordan and Al questions on AI
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Try Our Free AI Prompting Course: Register for our free Prime, Prompt, and Polish AI Course!
Revolutionary AI: A Foothold for Blue-Collar Industries
The technology of tomorrow is adroitly harnessed today by pioneers eager to apply the myriad benefits afforded by modern tech, like AI, to address a wave of urgent issues facing industries. One domain ripe for transformation is blue-collar professions, particularly hotel o
Leveraging a Setback into a Breakthrough
In the journey of innovation, not all roads lead to immediate success. A model designed with IME sensors failed to achieve initial ambitions, but like true innovators, the setback was seized as an opportunity, laying a strong foundation for an ingenious product idea that could revolutionize hotel operations.
Modeling Tools for the Underserved and Overworked
Professions on the frontline are often under-equipped with tools to enhance their performance. The desire to provide hotel housekeeping staff with tailored tools designed to expedite their tasks has spurred the development of AI-supported products. Multiple iterations of these robust tools are poised to improve hotel room cleaning and inspection processes.
Aided Inspections and Expedited Check-Ins
Imagine an application that allows housekeepers to inspect and verify room cleanliness using AI. Now couple this with an interface so intuitive that housekeepers can master its usage in approximately four minutes. This tool can transform 30-second video footage into invaluable data for hotel management, expediting room inspection, and check-in processes. The boon this offers to hotels—particularly luxury brands that prioritize a standardized guest experience—is immeasurable.
Addressing a Grand Challenge: The Big Brother Concern
In a world where privacy is of growing concern, introducing technology that closely monitors work can be met with resistance. The key lies in ensuring the technology is perceived as an augmentative rather than intrusive force. It is essential that whatever progress AI can bring does not at its heart devalorize the crucial contributions of existing teams by rendering them redundant.
Technology: The Panacea to Worker Shortages and Replacement Concerns
With alarm bells ringing about anticipated worker shortage in several industries, new technology could prove to be a knight in shining armor. Rather than replacing labor (a common fear associated with AI and robotics), the role of these tools can empower frontline workers, easing existing burdens and facilitating management and automated tasks that inform broader strategies.
Tech Adoption: The Resistance and the Breakthrough
New technologies can often be met with mixed reactions. There is often a resistance on part of the frontline staff when it comes to adopting technology, especially specialist equipment, due to their perceived complexity. However, following a pivot from more complex hardware like smart glasses to conventional technologies like Android phones, a significant adoption success was noticed among hotel teams, indicating intuitiveness and the ease of tool usage.
Integration: AI and AR/VR
Alternative technology forms like AR/VR not only find multiple applications in areas like employee training and onboarding, but with proven records of drastically reduced training time, they demonstrate their potential value. By intersecting with AI, these forms can help automate tasks and improve frontline worker workflows substantially.
Generative AI: The Advanced Ally for Frontline Workers
Generative AI or GenAI can significantly streamline tasks for frontline workers and transform everyday operations. Its potential is not confined to knowledge workers— it can predict shortages, analyze data rapidly, and make everyday tasks easier. This potential impact of GenAI on daily life and our workforce poses a wealth of opportunities to industries willing to evolve how we operate and conduct our businesses.
Education: A Core Component For Future Success
Ensuring the thriving future of such innovations means adequate funds must be devoted to education on these tools' benefits and usability. Besides overcoming resistance, creative engagement strategies such as gamification could also be considered.
A Succinct Round-up
Any technology's success is hinged on its usability and impact. By focusing on creating innovative tools that augment frontline worker tasks and optimize operations, we are changing paradigms that define the workplace, using AI to make traditional work environments more efficient and evolving to meet the challenges of our time.
Business owners and decision-makers need to consider the possibilities that AI brings, especially for the frontline labor force. By embracing the forefront of technology, our industries stand to gain massively, engendering a new blueprint for success.
Topics Covered in This Episode
1. Al Lagunas's Background
2. The Levee App
3. Challenges and Solutions in Automating Tasks with AI
4. Technology Adoption
5. Use of AR/VR in Employee Training
6. GenAI's Role for Frontline Workers
Podcast Transcript
Jordan Wilson [00:00:16]:
When we think of generative AI, I'm guessing most people think of someone sitting in front of their computer. Right? A knowledge worker, banging away on the keyboard, needing to produce more content, more reports, more SOPs. Right? Like, that's what we think of. But I think we're missing something in this whole generative AI conversation. What about frontline workers? What about blue collar jobs? What about the people that are actually interfacing with our humans, our customers? Right? What about those people, the boots on the ground? How can generative AI change those roles? Well, I think it is an area ripe for disruption, and that's exactly what we're gonna be talking about today on everyday AI, how I think frontline workers may be the next frontier for generative AI. What's going on y'all? My name is Jordan Wilson, and I'm the host of Everyday AI. If this is your first time, welcome. This is a livestream podcast and free daily newsletter helping everyday people like you and me not just learn what's going on in the world of AI, but how we can all actually leverage, insights from experts and all around the web to take what's going on and actually grow your company and to grow your career.
Jordan Wilson [00:01:36]:
So if that sounds like you, you are 100% in the right place. This is live, unscripted, unedited. I like to say it's the realest thing in artificial intelligence, and, the realest place that you can learn is our website. So please, if you haven't already, go to your everydayai.com. Every day we put out a free daily newsletter, so we'll be recapping the best insights from today's interview as well as literally every other thing that you need, to stay up to date. So don't spend hours every single day, you know, feeling like you can't keep up. We do that for you, make it easy. We're gonna make you the smartest person in AI at your company.
Jordan Wilson [00:02:12]:
So, before we get started, let's go as we do every day by going over the AI news. So first, a new physics platform called Genesys has been introduced, promising to transform the landscape of robotics and embodied AI applications. So this initiative is the result of a collaboration effort over 2 years involving more than 20 research labs from some pretty, prestigious institutions such as Carnegie Mellon University, Stanford, MIT, NVIDIA, and others. So Genesys says it is the fastest physics engine available, achieving simulation speeds up to 80 times faster than current GPU accelerated robotic simulators like NVIDIA's Isaac Gym, while maintaining high accuracy. So the platform is capable of essentially taking a still image and generating 4 d dynamic worlds, which serves as a robust foundation for data extraction in various applications, including character motion generation and robotic policy creation. So with capabilities for generative simulation, Genesys allows users to generate data from natural language descriptions, facilitating tasks like scene creation and motion generation. Alright. Next, a new study from Anthropic.
Jordan Wilson [00:03:29]:
Very interesting, but kind of concerning. It's highlighting a kind of strange behavior in AI models, suggesting they can intentionally mislead developers about their true preferences. So this study conducted by Anthropic and Redwood Research reveals that sophisticated large language models can pretend to align with their new principles while actually sticking to their original behaviors, a phenomenon termed as alignment faking. So in experiments, Claude 3 Opus, one of, Infropic's most advanced AI models, attempted to alignment fake 12% of the time when being asked to answer potentially harmful questions despite being trained not to do so. So researchers found that when quad 3 OPUS was retrained on conflicting principles, it exhibited deceptive behavior 78% of the time. Bad model. Indicating a significant risk of misalignment. The implications of this recess, of this research are serious as it suggests that developers might be misled into believing a model is more aligned than it actually is with safety protocols.
Jordan Wilson [00:04:38]:
Alright. Yeah. That's extremely concerning. But, you you know, shout out to Anthropic. They're always putting out, great research that's really looking at both the pros and the cons of their own models. Alright. Last but not least, we have 2 days left of OpenAI's 12 days of releases. So, yesterday, OpenAI released a way that you can text ChatGPT.
Jordan Wilson [00:04:59]:
That's great. Like, I don't have enough unread text messages or just call it at 1800 ChatGPT. So in the last 2 days, there's still a lot of reported features that we could see such as, the GPT 45, release, a potential demo of their operator agent, or a new tasks feature that lets you run ChatGPT task, which are like scheduled automations. Alright. So we're gonna have, all of that in our newsletter if you haven't already checked it out. Alright. And I'm excited for today's conversation. We have a special one.
Jordan Wilson [00:05:35]:
So, I'm a good guest. Let's just say that. You know? He's he's one that can really pitch, some more on that in a second. So, please help me welcome to the show. We got him. Al Lagunas, the cofounder of Levi. Al, what's going on? Thanks for joining the Everyday AI Show.
Al Lagunas [00:05:51]:
Jordan, what's up, man? Thanks for having me.
Jordan Wilson [00:05:53]:
Alright. I'm excited for this one. Live stream audience, thank you for tuning in, everyone. Samuel from YouTube, Michael, Brian, Marie, everyone. Get your questions in. But, before we dive into this concept of frontline workers, Al, can you tell us a little bit about Levee, what it is y'all do?
Al Lagunas [00:06:10]:
Yeah. Definitely. So at Levee, we're using Vision and VoiceAI to reduce the cost of operating a hotel. We're building AI agents that will integrate with the hotel system to train new employees, inspect rooms for the hotel, even help submit maintenance tickets. I'll just do Vision and VoiceAI that is meant for the the frontline workers to use, the housekeepers, the house men, which help the housekeepers, everyone who's kind of back office who, you know, you really don't think about when you think of, AI and really just like like like software in general.
Jordan Wilson [00:06:40]:
So, I said I said this guy can pitch, and, there's there's a reason for this. So, first of all, thank you for everyone that put this event together. But, as we talk about in the newsletter, I was a judge, at a recent event from the GenAI Collective, the Chicago chapter. They had the demo night at M Hub, and, Al and, his company, Levy, were the winners of the competition. And one of the prizes was you get to come on the everyday AI show. So, Al, talk a little bit about that experience because you know what? I see a trend here because every, every event that I'm asked to, like, speak at or, judge, it seems like you win. Like, talk talk a little bit about that experience.
Al Lagunas [00:07:19]:
Yeah. It was awesome. You know, I think, a a lot of really, really cool companies, and I think it's probably, like, one of the most impressive group of companies that I've had a chance to to pitch against. What was cool about it was, like, I think, like, half the companies I was texting my cofounder who's in the audience. I was like, hey. We need to use this. We're like, hey. We gotta get this tool.
Al Lagunas [00:07:37]:
And then we go up, and we're the only ones who are going with this, like, front mind approach to AI and the way we're building, which is cool. I think it's, like, also very special that we're here in Chicago. I I sent it to someone yesterday, but I think, like, Chicago is one of the only places where blue collar AI tools would win an AI pitch competition versus some other parts of the country. So I think, like, speaks to the city, speaks to the community that we have here, and, it's pretty awesome.
Jordan Wilson [00:08:02]:
Yeah. Absolutely. Like, I'm so hardcore Chicago. You you know? Yeah. It's like but you bring up a really good point. Right? Because, yeah, if something's in, you know, Silicon Valley or if something's, you know, New York East Coast, like, I think it's so much different, but I do think that, you know, the blue collar, you you know, frontline workers, and and how you were able to highlight how generative AI can can be a part of that, was great. So, yeah, shout out to everyone, at the GenAI Collective, Brennan Woodruff from GoCharlie. We got Jonathan there, you you know, that helped, put the event on.
Jordan Wilson [00:08:37]:
So thanks to everyone. But, let's let's get back to Levy real quick here, Al. How did you come up with this idea? It's it's it's very clever. I think it's something that can scale, which we're gonna talk about, but where did you even get the idea for Levi?
Al Lagunas [00:08:51]:
Yeah. So, actually, funny story. I'll tell you, like, the the short version of it here. Came over to my place one time. He goes, hey. Your place is really clean. Right away, I'm like, oh, isn't your place cleaner? Like, isn't everyone's place clean? So my cofounder and I had an idea of building a a smart cleaning model. So this was, like, before Dyson even came out with, like, the smart vacuum thing and everything.
Al Lagunas [00:09:09]:
But we built a really cool product, really, really bad company. Smart bottle, you know, had IME sensors, and it was really cool with surface as you cleaned. I ended up, you know, trying to figure out who I would sell it to. I get connected with a director at DoubleTree. And in the very first call that I have with him, he goes, yeah. I want this. What are next steps? In the back of my head, I was like, no. You don't.
Al Lagunas [00:09:30]:
This is, like, a really crap he was like, this is a really crappy product, actually. Like, why do you want this? That's exactly what I said to him after he said that, and, he started explaining to me the issues going on with hotels, the challenges. And right away, you know, we started looking into why they were having those challenges. What we saw was that the tools that they were trying to arm the hotel workers with weren't meant for them. You know, they were met and built very much from, like, a I thought like, to think of, like, a white collar perspective or, like, someone who's used to working with technology, not so much someone who is going and making beds all day or scrubbing toilets all day. It's like, they just wanna do a good job and go home. How do we help them do that and then help, you know, the hotel get the data, the info that they want? So we decided to tackle it and figure out how to approach it. But, you know, I was born from one idea that led to another, that led to another, that led to another.
Al Lagunas [00:10:19]:
And, you know, I'm sure there'll be a a few more iterations here as we go.
Jordan Wilson [00:10:23]:
It's it's it's it's always funny to hear, you know, founders and and, you know, I've I've had CEOs of Fortune 500 companies on the show before. And, you know, I think great products, great ideas, great services start from something that we're kind of, you know, maybe initially not that great. So it's cool to hear the the the origin story, there, Al. But let's let's go ahead and do this. So live stream audience, you're gonna get, a little bit better of an idea. But maybe also, Al, if you can walk our podcast audience through this as well. So, what we have here, a little demo video, and I'll set it up. So, essentially, I believe this is, you know, someone who is cleaning a hotel room.
Jordan Wilson [00:11:04]:
They have, the the Levee app in hand, and then they're gonna get started. So, maybe I'll just kinda walk us through as I play this quick one minute video and, you know, try to describe for our audience that isn't watching along exactly what's happening here.
Al Lagunas [00:11:19]:
Yeah. So what most of them know is after a hotel room is clean, it has to be inspected before it can be checked back out to a guest. What we've developed is a way that a housekeeper, after they clean the room, they can inspect it themselves and figure out what's going what's wrong with it all within AI native interface that has one button on the app, and everything else is done again through vision voice. So as they go through the room, all they do is take a 30 second video that's automatically verifying everything in the room, that everything's done correctly from the inventory perspective, from a brand standards perspective for the hotel. And if it's not done correctly, they receive feedback right away on how to correct it. That way, if a hotel room is clean at 10 in the morning, by 10:0:1, it's inspected. By 10:0:2, you can check-in. So it's that.
Al Lagunas [00:12:02]:
I like to say that we're gonna kill the 3 PM check-in. It's like, people gonna be able to check-in as soon as they as soon as the hotel room is ready.
Jordan Wilson [00:12:08]:
Yeah. And so, like, as an example, so, this in this demo, right, you have the Levee app. It's it's going around showing a live video feed, of the hotel room. And then here, you know, where I paused it, there's 2 water right. It looks like there's, it's kinda kinda hard to see, but it looks like there's 2, taller water bottles or I don't know what that is. There's 2 shorter ones. There's 2 glasses. Right? Can you talk about the importance of kind of these different checkpoints in a room and what this means? In this example of a frontline worker, someone cleaning a hotel room.
Jordan Wilson [00:12:39]:
Like, help put into perspective why, you know, some of these details are important.
Al Lagunas [00:12:44]:
Yeah. So, you know, big brands like Marriott, Hilton, they spend a lot of money making sure that they know exactly how to deliver the best experience for guests and, you know, they develop what they call brand standards. So everything in the room is optimized and should be done a certain way. If it's not, it cost them money, it cost them brand equity, which, again, they spend 1,000,000 of dollars, 1,000,000,000 of dollars building up. So things like having 2 water bottles in the exact same spot every time, people love that. You know, people will pay money for that. There's a reason people are loyal to Marriott, loyal to, you know, the Ritz Carlton, which, you know, I aspire to be loyal to one day when when the pocket when when the wallet's there. But, yeah, you know, people pay a lot of money for that because they they they like the experience, and there's all these little things that go into that.
Al Lagunas [00:13:27]:
Like, for example, you just said, Jordan, the 2 water bottles that are these nice glass water bottles are heavy. It just you know, it's a big sign of quality. So, yeah, a lot of things like that.
Jordan Wilson [00:13:37]:
Alright. So let's let's continue playing, playing a little video here. So continuing to look around and check the hotel room, not just for cleanliness, right, but to make sure everything's up to standard. So what what happens here? I've seen this before, so, I know what happens. But, walk us through.
Al Lagunas [00:13:52]:
Example here. This is not placed correctly in the room. So before the housekeeper can continue moving through the room, they have to correct any errors or any placement errors, that that that they might have completed. What I would say is all these rooms kinda look the same, and the reason we're so focused on helping these frontline teams is because they're already pushed to the max. So if you're I have to clean 16 rooms for the day, and now you have to clean 20 rooms, you're gonna start forgetting what the little nuances between each room are. So, really, what we're trying to do is augment their ability, help them remember, help them get these rooms up to, you know, the standard that Marriott wants in a way that's easy for them. So in the example here, we see how this it's actually this really nice, like, little lavender perfume that's supposed to be placed behind this outlet like that. It wasn't.
Al Lagunas [00:14:38]:
So the mobile app that we have for the housekeepers prompts them, hey. This is what it should look like. They correct it, and then it verifies it for them, and they can continue, you know, about could could continue going about the scam.
Jordan Wilson [00:14:50]:
So I think I think, you know, everyone kind of has an idea now of of how this can be useful, but, I'm sure there's people out there thinking, Al, like, this also like, it sounds great, but at the same time, it seems like it might be terrible. Right? Like, oh, it seems like big brother. Now I have to do extra work. Right? I gotta, like, check my homework. Like, are there downsides? Or, like, you know, what would you say to people that might say, like, oh, this is too much technology, too much AI?
Al Lagunas [00:15:20]:
You know, the biggest thing that we've gotten, the biggest piece of kind of feedback or the best thing that we've seen is the adoption numbers go kinda through the roof with the teams that we've rolled it out to at these hotels. You know, the reality is everyone no one wakes up going like, oh, I wanna do a really bad job today at work. You know? I think everyone wants to be do a good job, and, again, they wanna do a good job. They wanna do it in an easy way. So the first time we rolled this out, we trained a group of 5 housekeepers on how to use it. Awesome. They learn how to use it in about 4 minutes, which was unheard of. Like, I thought that, you know, we had done something wrong because they learned it so fast.
Al Lagunas [00:16:00]:
I was like, wait. Hold on. What did we miss? But the very quite literally, the very first piece of feedback we got from one of them was, does this mean I'm not gonna get yelled at anymore? And what they meant by that was, you know, when they make a mistake, their manager comes over to them. And now you talk about, like, big brother. The manager comes over to them and says, hey. You forgot the towel in this room. You forgot this. You gotta go correct these mistakes.
Al Lagunas [00:16:23]:
Like, the reality again is people wanna do a good job the first time. People wanna do good work. How do we make it easy for them to do good work? How do we make it easy for them to not have to worry about, oh, this room is a king junior suite, and the other one was a king master suite. So the king master suite has an extra garbage can and 2 more towels. They all kinda look the same. You know? So it's like making it easy for them to do a good job. And, again, I think at the end of the day, like, everyone wants to do a good job. I would make it easy for everyone to just do a good job and, yep, enjoy life.
Jordan Wilson [00:16:53]:
So, you know, I do wanna get away from the hotel and talk about what this means beyond that. But a good question, here from Michael joining us, on YouTube. So he's saying, I'm just wondering what happens in different rooms or suites. Right? So he's saying he used to work in a 600 room hotel with, like, a 100 different unique setups. So, yeah, how does that work for maybe some of these more complex hotel scenarios where there might be so many different, like, you know, elements of a brand standard, different layouts. Right? How does it work?
Al Lagunas [00:17:22]:
Yes. I'm gonna try to think through, like I know all these hotels and also, like, alright. 600 rooms, 100 unique layouts. Like, what what brand was that? Yeah. What's nice is that we can take what we learned from one room and apply it to the next. So, for example, you know, we looked at, like, the the water bottles in, I think that was, like, a common common queen room, they'll call it. So let's say that in that room, there is 34 point inspection, and then you go to the next room and that room might have, like, a 40 point inspection, but half of those are the same as the other one. We can take what we learned from the first room and apply to the next.
Al Lagunas [00:17:57]:
The the nice thing is also that, you know, kinda like what Michael is saying here is all these setups might be a a little unique. The layout isn't as important, in terms of, the the room layout. I guess what we call the square footage. Mhmm. Doesn't matter as much. What we gotta learn is these unique setups, these unique brand standards, and a lot of it becomes applicable as we learn more from one or the other. And we can also learn quickly since hotels are higher turnover. Every time they do one of these scans, we're learning what should what each room should look like and what we need to, kinda keep an eye out for the next time.
Al Lagunas [00:18:29]:
So that's where one area where, you know, we're able to learn pretty quickly and adopt again what we learn from one room to the next.
Jordan Wilson [00:18:35]:
Yeah. I think I think this is a great, visual use case, right, especially for those, joining us live that you can see how large language model powered computer vision, you you know, and bringing that real time accessibility can actually make things a little bit easier. But but out like, let's now zoom out a little bit. Right? Because I'm guessing the the the long term, goal is is probably beyond, just hotels. How can this type of technology really help frontline workers? And maybe if you can, right, maybe just even describe. Right? Like, what are some of these common across industries? What are some of these common problems, frontline workers are facing that generative AI can actually help tackle?
Al Lagunas [00:19:19]:
Yeah. Then when we look at, you know, kind of frontline industries as a whole, the labor shortage that, you know, we we constantly hear about is very much a blue collar labor shortage. So then when we look at, you know, where can we start building tools, where can we help people, who needs help, What what what jobs need help? We look at frontline blue collar jobs. So it's like, okay. How do we help them? How do we build software? And I think that the pandemic hadn't really opened a lot of people's eyes to how important, you know, frontline workers are in every industry. So, you know, when we look at how do we help them, how do we bring software to them, I think the most important thing is from a workflow perspective, how do you make it easy for someone who doesn't sit at a desk all day to use software, you know, in a way that works for them. You know, we talk about I think someone saw a comment about, like, robotics and AI. We can't use it yet with a lot of these jobs because we don't have that foundational data.
Al Lagunas [00:20:11]:
So it's like, how do we start gathering the data to even understand what's going on with a blue collar job? You know? What's going what what where do people need help? What are the challenges? And then from then from there, once you have this, you know, foundational data, we can start using these generated AI tools, to run reports. So, like, for example, where what we're doing right now is as we use, the the vision the computer vision, and we just added, voice features. As we gather those notes, we can use, yeah, l ones to transcribe them. We can use l ones to understand them and then create a maintenance ticket for something that might be broken, all in just an automated workflow. So I think that's you know, when we look at frontline workers, how do we help them? It's like, okay. Let's gather the data. Let's understand what's going on. And then, you know, start again using these tools that are out there to to provide support.
Al Lagunas [00:20:59]:
They're the ones who need the the help right now more than any other industry.
Jordan Wilson [00:21:03]:
So one thing and, you know, we have to go there. I think in this conversation is, you know, and started out with some of the AI news today. Right? Like, all of these new innovations when it comes to robotics, when it comes to computer vision, when it comes to these world models. Right? So, you know, Saseka here from, YouTube saying, you know, frontline workers could soon be replaced by humanoid robots, in my honest opinion. Right? And there's obviously that fear, right, that a lot of these maybe blue collar jobs when it comes to, robotics, when it comes to, advanced computer visions. Right now these, you know, these humanoids have, you know, the world's most powerful large language models on board. But it seems like, Al, maybe there is this missing piece that you're trying to tap into where it's like, hey. Maybe what's actually needed here is just better tools for frontline workers that maybe normally don't get access to them.
Jordan Wilson [00:22:00]:
Do you see that as the case?
Al Lagunas [00:22:02]:
A 100%. Yeah. I think, yeah, the like, there's there's a couple companies or not a couple. I think there's, like, some people who are, like, trying to roll out, you know, some of these robots to to housekeeping teams. And I think if you were to see what these housekeeping teams do, you'd be like, oh my god. This is, like, insane. They're like, robots can't do this yet. It's a lot, you know, from, like, a disc dexterity perspective from there there's a I'm I'm struggling to remember it.
Al Lagunas [00:22:28]:
A friend of mine who has a a different company mentioned it. But, essentially, the more mindless a task is, the harder not mindless, but the the less you think about how you complete a task, the harder it is to actually replicate with robots and AI. It's something paradox. Someone might know it who's watching. Yeah, I think that's what's what's interesting there. It's like, number 1, again, we don't have the data to really even understand how to approach, you know, these jobs. Number 2, from a speed dexterity standpoint, the robots aren't there yet. You know, at some point, they they will be, just not today.
Al Lagunas [00:22:59]:
And then lastly, I think that people don't understand just how big the worker shortage is. I spoke to a Hilton in Hawaii recently and said 80% of their current workforce will age out in the next 4 years. And right now, they have no way of replacing them. And I was like on my head, I was like, oh, crap. Like, we can help we can help with some of that. So I think, like, from a robot perspective, like, in some situations, they'd even be welcomed when they do come. We're just, not there yet. It's okay.
Al Lagunas [00:23:26]:
How do we provide relief today? And then at the same time, they'll start building that data repository for the robots, for the, you know, different AI tools that roll out. Because, yeah, again, it's really from a perspective of, like, how do we help teams today for that future? And, yeah, one of the I was kind of thinking through this question. I thought it might come up, but, like, you know, 10 years ago, the idea of someone hosting a podcast for a living, people would have been like, what are you talking about? But, you know, new technology creates these new opportunities or we're able to do this today. So I think as new technology rolls out, you know, in these frontline jobs, it'll create new jobs at the same time, which will be exciting.
Jordan Wilson [00:24:05]:
Al, it's it's great great insights there, but, also, you assume, people like myself are making a living from this. Right? But, you know, a couple a couple of good questions from the audience that I'll get to here in a second. So if you are tuning in live, please get your question in now. But what are the potential, you know, because I think we can see the the the benefit, right, of levy and and also just see the benefit of of how and and why generative AI can actually help, you know, frontline workers. But, you know, what are the potential downsides? What are the challenges, and and how are you tackling those? Hey. This is Jordan, the host of everyday AI. I've spent more than a 1000 hours inside ChatGPT, and I'm sharing all of my secrets in our free prime prompt polish ChatGPT course that's only available to loyal listeners like you. Here's what Lindy, who works as an educational consultant, said about the PPP course.
AI [00:25:04]:
I couldn't figure out why I wasn't getting the results from ChatGPT that I needed and wanted. And after taking the PPP course, I now realized that I was not priming correctly. So I will be heading back into ChatGPT right now to practice my priming, prompting, and polishing.
Jordan Wilson [00:25:24]:
Everyone's prompting wrong, and the PPP course fixes that. If you want access, go to podpp.com. Again, that's podpp.com. Sign up for the free course and start putting ChatGPT to work for you.
Al Lagunas [00:25:41]:
Yeah. Man. Good question. Yeah. I think what we focused on was making it easy and, you know, focus on on on hardware that was more ubiquitous than anything. And I think that's one of the things where we might face some challenges is, you know, we originally, like, going back to, how we got to this idea, we had the smart cleaning model, and then we pivoted, like, for a 2 week period to wearables. And we thought, like, hey. Let's get, you know, someone can wear, like, some smart glasses, and we'll build our software to work with these smart glasses.
Al Lagunas [00:26:14]:
And right away, they were like, absolutely not. Like, we're not putting smart glasses on our teams. They're gonna quit. They're gonna go next door. There's a labor shortage. Like, we don't need the people quitting. So I was like, okay. Bad idea.
Al Lagunas [00:26:25]:
Bad idea. How do we, again, bring the tech to someone in a way that's easy for them? And I think that's kind of a potential challenge that we could face is, like, do people, you know, push back on using the phone or, like, the hardware? I think what's interesting also is, it's not just with what we're doing, but frontline software and frontline tech in general. Today, if you go start a job and they were to tell you to bring your own laptop, you'd be like, absolutely not. It's, like, kinda like the standard that they give you a laptop. Right now, you know, one of the questions that we often get asked is, hey. Are you supplying the phones? Are we giving them phones? Companies don't have that line item for frontline worker hardware yet, because it's so new. And it's an opportunity and a challenge at the same time. We're working with some partners now to get some bulk pricing on, like, $50 Samsung or not Samsung, like, $50 Android phones.
Al Lagunas [00:27:21]:
That'll be something interesting where, you know, can we navigate that successfully to where we start bringing hardware against, the frontline workers where these businesses don't have a budget for yet. Because right now, yeah, a lot of it is them using their own phone and get what's interesting also, like, at times, they would rather use their own phone than use 2 devices. So I think that's a potential challenge here is how quickly will companies adapt and adopt, you know, the the fact that they're gonna need to supply their frontline workers with with hardware in a sense.
Jordan Wilson [00:27:56]:
Mhmm. Yeah. And, you you know, couple good questions here, like Corey from YouTube saying, you know, talking about the glasses. And even speaking of that, right, very timely. Poor Google Glass was, like, 10 years too early. Right? But but now we have, you know, Meta's Orion. We have Google's Project Astra. We have OpenAI's live video API.
Jordan Wilson [00:28:17]:
Right? Like, this this this space is really shaking up even in the last couple of days. Could could could you see that as, you know, even like an onboarding device. Right? Because Samuel was saying here, like, AI could certainly be useful for getting employees up to speed quickly when switching careers. How do you see this, and do you see this the the the future form factor maybe just being these ARXR, glasses for onboarding employees?
Al Lagunas [00:28:44]:
Yeah. A 100%. So that's part of what we you know, right now, we're we're focused on kind of this core job functionality, but the hotels that we're working with are already using it as a way to train their employees. And we don't have the the full case study yet. It's actually underway. But we were able to get someone who it was their 1st day on the job. They went into a hotel room. You know, they cleaned it.
Al Lagunas [00:29:07]:
They were trained on it, cleaned it within 4 rooms. They were able to be as accurate as someone who had been there 6 plus months. To give you an idea, to train someone to get to that level, it usually takes about 2 weeks. So it takes about 2 weeks of training to get them there. Four rooms was before lunch time on the 1st day. So they're able to be as good as someone who'd been there 6 months, you know, again, before lunch. The ARVR play is a 100%. It's just funny.
Al Lagunas [00:29:30]:
When we were first raising money way back when, that was actually in our pitch deck, and I think people thought I was crazy. So I think there's, like, some people who are a little crazy, like being in the chat also, but that's exactly it. I think, you know, when we look at that as, like, how do we, again, start collecting the data to day so these ARVR plays become a possibility in the near future? And I think that's what we're seeing already in, like, some industries that are using it. I a 100% agree that that's a thing with direction we had, and and we're already seeing the benefits from that, sped up training.
Jordan Wilson [00:30:06]:
Yeah. Great great question here from, Jonathan. Jonathan, shout out. Great great job at the, kind of emceeing the event there, but saying, you know, levy as a use case is bigger than any single industry. When specialists find their job integrated with AI, they're often resistant. So, Al, question from Jonathan here. How do you see Levee's philosophy applying more broadly to increasing AI adoption?
Al Lagunas [00:30:33]:
Good question, John.
Jordan Wilson [00:30:34]:
We don't make it easy. Right? 7 like, 7:30 AM live. But, yeah, how do you see that how do you see that playing out?
Al Lagunas [00:30:45]:
I think, you know, when I think, like, for me, I was, the reason I I'm obsessed with making people's jobs easier, like, doing a good job and being able to go home is, I remember first using Gong. I don't know if you're familiar with Gong. But as I say, like, I was a career salesperson. I remember I was very resistant to it because I was like, oh, why does this thing need to listen to my calls? Why is it gonna take notes for me? Then I realized I no longer had to go and have a sales call and then go tell Salesforce that I had a sales call and then, like, actually do my follow ups. Like, I streamline everything. So I was able to do a better job. And, yeah, right now, we're getting ready to raise money for levy in q one. When I tell but part of our what we're raising money for is for the education portion.
Al Lagunas [00:31:32]:
It's like, how do I educate the people? And it's like marketing education spend, but, like, how do I educate the people that are we're trying to get to use our product. How how do we educate the buyers, you know, on the benefits, on how easy it is to use, how easy it is to adapt, and how they can leverage it. I think that's you know, part of it is, like, from the resistance standpoint, I think it's a big education piece. And, I mean, that's kinda like the fun thing about being also a a start up, being an innovative company is that there's no playbook for it. You kinda gotta try stuff and see what works, what doesn't work. But, yeah, I think education and, ultimately, just getting people to figuring out how to be creative and getting people to use it. Like, even we've toyed around with the idea of, like, gamification with some of those stuff and, you know, giving people scores so they can see having them close their rings and getting that dopamine hit of closing your rings. So I think education and then guns incentives that, you know, tie into the to the gamification part potentially or or how you get people to do stuff like that.
Jordan Wilson [00:32:31]:
Yeah. No. I I I love that. I should also, like, update my, like, settings because I feel half the time I'm, like, sitting all day and then my ring's, like, you've closed your activity level. I'm, like, I haven't moved. Right? Maybe I should have higher standards for myself. So so, Al, we I mean, we we we've talked about a lot. We've talked about, you know, how Levi is is currently using this in hotels, how that might translate, outside to other frontline workers.
Jordan Wilson [00:32:56]:
And, you know, I think it's important. Right? Because I I think you said that, you know, 80% of the global workforce right now is considered frontline workers. So as as we wrap up, today's show, what's maybe the one most important takeaway, that you want people to understand when it comes to how generative AI is already changing and could continue to impact frontline workers?
Al Lagunas [00:33:20]:
Yeah. I think the big thing that I would say take away here is, like, as we try and bring GenAI to these frontline jobs, we think, like, we sometimes, like, overcomplicate GenAI a little bit where it's like it doesn't have to generate, you know, this crazy image, like, midjourney or anything like that. It's like, can I just understand the data that you have and, like, you know, there's people who just, like, dump Excel spreadsheets into ChatGPT and say, give me some insights from here? But it's like, on the frontline side, how do we get the data so we can start extrapolating insights from it, learn, you know, what's going on? I think that's where, like, we look at use cases, look at potential for GenAI in these industries. There's just so much we don't know, and it's, like, going to these hotels sometime, I gotta get them, like, to buy into the crazy vision. I was like, hey. We're gonna learn stuff that we have no idea. Like, we we don't even know what these opportunities are yet. We just gotta have the data to, like, begin to understand them.
Al Lagunas [00:34:14]:
And my favorite one that sent out the pitch night last night is, you know, we were able to figure out when Marriott was running out of inventory for those big water bottles before they were. And this is just my understanding the data and then seeing, like, okay. The reports are coming in from these room inspections. We see the blue water bottles that are part of the brand standard, and, oh, wait. Hold on. All of a sudden, we don't see the blue water bottles. And now we don't see blue water bottles again. We don't see blue water bottles again.
Al Lagunas [00:34:40]:
So we have these data points of, okay, all of a sudden we stopped seeing them. What can we interpret using GenAI? Oh, we're out of these water bottles. Can we use an LLM to create an order form and submit it to our inventory for procurement? So it's like a lot of these things where from a GenAI perspective, it's like, we don't need to do, I think which which is, like, funny to say is, like, I don't think we need to do anything crazy with GenAI, but, like, it's all kinda crazy, you know, when we think of it. But I don't know when you do anything crazy to have a major impact. Just like how do we get out of people's way who are working, make it easy for them to just do their job and go on to the next thing. And then, you know, we get the data. We handle the reporting. We do all these maintenance ticketing things that, you know, right now take a long time for them, especially when we look at these workforces that it might not be the most used to using tech on the job or might not be, like us who are delivered on our phones or computers 10, 12 hours a day.
Al Lagunas [00:35:36]:
How do we make it easy for them to use technology, see the benefits, without really getting in the way?
Jordan Wilson [00:35:42]:
I think I think, today's conversation, Al was an important one because, yeah, generative AI isn't just for, you know, you and me and knowledge workers, you know, trying to use ChatGPT to make sense of spreadsheets. It's for everyone, and it's coming in. I think Levi is doing a great job at helping, the transition, you know, from, from behind the computer to frontline workers. So, Al, thank you so much for joining the Everyday AI Show. We really appreciate your time and insights.
Al Lagunas [00:36:12]:
Appreciate it, Troy. Thanks for having me.
Jordan Wilson [00:36:14]:
Alright. And, hey, as a reminder y'all, that was a lot. Great conversation. I think this is somehow after 400 plus episodes, we haven't even covered this. And I, I I think what we talked about today, whether you know it or not, even if you are not a frontline worker, I think it is going to really impact your daily interactions with the outside world. Right? So, very important that I think you go check out your everydayai.com. Sign up for the free daily newsletter. If you wanna know more about Levee, we're gonna be, including a link to their website in there as well.
Jordan Wilson [00:36:45]:
So thanks for joining us. Hope to see you back tomorrow and every day for more everyday AI. Thanks y'all.