Earley AI Podcast – Episode 51: AI Startup Agility and Workforce Digitization with Jason Radisson

Guest: Jason Radisson, Founder and CEO at Movo

Hosts: Seth Earley, CEO at Earley Information Science

             Chris Featherstone, Sr. Director of AI/Data Product/Program Management at Salesforce

Published on: July 31, 2024

In this episode, Seth Earley and Chris Featherstone speak with Jason Radisson, a digital transformation expert who has led initiatives at McKinsey, eBay, and Uber. They explore misconceptions about AI startup capabilities, how enterprises can effectively partner with agile teams, and the lessons learned from digitizing frontline workforces in the gig economy. Jason shares insights on automation versus change management, the fallacy of starting with enterprise data warehouses, and how marketplace efficiency principles from companies like Uber can transform traditional workforce management across industries like logistics, healthcare, and hospitality.

Key Takeaways:

• Quality AI startups with experienced enterprise teams can deliver new features and integrations in weeks rather than months, contrary to corporate expectations about startup capabilities.
• True automation obviates work rather than requiring year-long change management processes, representing a paradigm shift from the ERP optimization era that many enterprises haven't adapted to.
• The fallacy of starting with enterprise data warehouses wastes resources—successful implementation requires working backwards from customer-first use cases that drive immediate business value.
• Gig economy marketplace principles can digitize traditional frontline workforces through machine learning-powered scheduling, task distribution, and real-time optimization serving millions of workers across industries.
• Organizations must clearly identify and quantify their top automation opportunities with specific dollar amounts before execution, as knowing the strategic imperative without organizational muscle fails.
• User experience becomes the ultimate differentiator when core functionality is comparable, as modern workers demand mobile-first interfaces that align with their mental models and workflow patterns.
• Successful workforce platforms create win-win scenarios by honoring employee preferences for most shifts while maintaining employer flexibility, avoiding both rigid scheduling and complete gigification extremes.

Insightful Quotes:

"True automation doesn't require much change management because it's usually obviating work, which is very different from the ERP era where you're mapping the company's processes and optimizing them maybe 10 or 20%. It's literally plug in the system and then turn off the team that used to do that work." - Jason Radisson

"One of the biggest fallacies in IT is that everything starts with an enterprise data warehouse and a data lake. It's very expensive, it's very self-serving for database vendors. You have to work backwards from the use cases that ring the cash register." - Jason Radisson

"Having the answer doesn't matter. I can publish articles and say this is the way it is and here's how you do it and here's the case study. And yet there's this weird gap. One of the biggest problems in corporate America is a lack of understanding where the big levers are." - Jason Radisson

Tune in to discover how startup agility combined with enterprise experience creates transformative AI solutions—and why execution muscle matters more than strategic knowledge when digitizing frontline operations at scale.

Links:
LinkedIn: https://www.linkedin.com/in/jason-radisson/
Website: https://www.movo.co


Ways to Tune In:
Earley AI Podcast: https://www.earley.com/earley-ai-podcast-home
Apple Podcast: https://podcasts.apple.com/podcast/id1586654770
Spotify: https://open.spotify.com/show/5nkcZvVYjHHj6wtBABqLbE?si=73cd5d5fc89f4781
iHeart Radio: https://www.iheart.com/podcast/269-earley-ai-podcast-87108370/
Stitcher: https://www.stitcher.com/show/earley-ai-podcast
Amazon Music: https://music.amazon.com/podcasts/18524b67-09cf-433f-82db-07b6213ad3ba/earley-ai-podcast
Buzzsprout: https://earleyai.buzzsprout.com/ 

Podcast Transcript: AI Startups, Enterprise Transformation, and Digitizing the Frontline Workforce

Transcript introduction

This transcript captures a conversation between Seth Earley, Chris Featherstone, and Jason Radisson about overcoming enterprise misconceptions regarding AI startups, the critical differences between automation and traditional change management, lessons from the gig economy applied to frontline workforce management, and why knowing strategic opportunities without execution capability leads to competitive failure.

Transcript

Seth Earley: Welcome to today's podcast. We're really excited to introduce our guest for today and we're going to talk about a lot of challenges and misconceptions around AI startups and AI startup capabilities, especially when it comes to large enterprises who may want to bring in an AI startup to help them with some of their challenges. We're going to talk about digitizing and managing the frontline workforce with AI and some lessons that have been learned from the gig economy. When you start thinking about all of the resources that need to be managed, you know, in, you know, a large city, when you talk about, you know, Uber or Lyft or other types of, or Doordash or other types of organizations where you have so much going on in the workforce, we'll talk about ways that get supplied to non gig economy types of organizations. We'll talk about user experience and how user experience is something that really makes a huge difference in terms of the ability to get AI successfully deployed in an organization and how critical that is to have that simpler user experience in order to take advantage of the capabilities that we're offering. So our guest today has successfully implemented and led digital transformations at companies like McKinsey, eBay and Uber. He's very passionate about leveraging technology in developing countries and empowering workers in the gig economy. Currently, he's spearheading a company that is focused on digitizing the frontline workforce to maximize productivity and efficiency. Jason Radisson, welcome to the show. Seth,

Jason Radisson: thanks for having me on. Thanks for the intro. So one

Seth Earley: of the things we like to kind of begin with is understanding some of the misconceptions in the space, especially the way you're working in organizations and you're looking at how to develop and deploy AI. And you know, a lot of large organizations, you know, number one, they can, they, they sometimes underestimate the agility that an AI startup can bring to the table. And the other piece of this is there's a lot, there's a lot of challenges around using startups because many times they don't necessarily have a deep track record and some of them can come and go. So why don't you talk a little bit about how organizations think about what are some of the misconceptions, how they look at AI startups and how they think about those challenges and how those, they can bring value to the organization. Yeah, absolutely.

Jason Radisson: Yeah, there are a couple of big ones. I think if you start from the top, one of the most important things, obviously I'll make the obvious point first, which is got to be quality, it's got to be a quality team. Got to be a team that has worked with enterprises before, doesn't need to be for 20 or 30 years, can be for two or three years, but rarely are you going to find a startup that is a couple of guys in a garage that have never seen the inside of a Fortune 500 company that are just going to come in and rock your socks. I think if you get beyond that and you're working with a good team, I think a lot of the corporate misconceptions, people just don't realize how quickly we work. They don't realize that in a small team with experienced engineers who have worked at all kinds of scales and all kinds of different problems, we can literally have a feature in a week or two, a completely new feature. If it's something as simple as an extra filter, a tweak, an adjustment, a quick integration. People talk about, oh my God, APIs and you know, how are we going to integrate this stuff from a quick moving startup that's an experienced team? We can get stuff done literally in a week or two. An integration might actually literally be some senior engineers time, like two hours if that thing is well specked out. And most of that'll be QA, the 15 minutes of code that he had to write or they had to write to get that done. So I think there's, there's a big misconception there and probably the third misconception is a lot of companies just don't know how to buy services from a startup. They don't know how to interface with them. You know, it's not sometimes RFPs, you know, the startup needs the feedback and sort of needs the, the close working relationship to be able to do things on the fly in real time and get them out and operating so that we can, we can adjust to what that big company needs. It's not an RFP thing. We're not sitting on top of 20 year old code, you know, and we're going to negotiate around some terms and spend three months doing that. All of that is a huge waste of time and value for the big enterprise. So those would be three, just kind of right off the top. But you know, there's a lot that we see out there where we just go, oh my God, this is, this is so difficult. And then, you know, there are, There are Fortune 500, there are Fortune 100 companies there know exactly how to buy from startups. You know, two weeks after you're talking with them, you're already rolling features out in some, you know, generally well protected, sort of low risk area of the business but you're already out there operating together. But you know, that's sort of a top 5 or 10% of enterprise companies and the rest really, really struggle working with a startup.

Chris Featherstone: You know what, I'm a super fan of agility. I've had a couple startups myself. The thing that's interesting and you've seen this probably and I'd love to get your take is when you walk into an enterprise organization and they say things like well, we don't know if we can take on that much change that quickly. Which is actually, I get it, it's a real fear. But how do you kind of get around that or help them, you know? Yeah, oh yeah, absolutely. And I think

Jason Radisson: what is, what is so just, just sort of paradigm shifting. Such a different world that we're in now is really true. Automation doesn't require much change management because it's usually obviating work which is very different from sort of the ERP era where you know, we're mapping the company's axis is, and we're adjusting them, we're optimizing them maybe 10 or 20%. And you know, all of that has to flow in this big long, you know, year long change management. All that stuff is pretty obsolete when you're talking true like machine learning, heavily automated systems. It's literally plug in the system and then turn off the team that used to do that work. And it's just, I think enterprise, non sort of tech companies just don't, they're not used to that kind of change management. I think it's very different. I would say like the Silicon Valley experience, the big tech experience is like things work that way. It's the opposite. You have this cultural pressure. And I'll sort of paraphrase what happens a lot in, in Silicon Valley. But you know, basically your manager comes to you and says hey, you know Chris, you've been doing that thing three times now. Why haven't you automated it? Why isn't engineering here like automating? Like what are you guys doing with you know, 30 guys working on something manually? That makes no sense. So there's, there's this bias in big tech to completely automate anything that is a stable process. Corporate America, you know, typical large enterprise company doesn't have that bias, doesn't have those muscles. And then you get to this, you know, machine learning is now, I think, you know, getting more sort of, at least sort of attention. And you know, we have this potential to do even more automation in white collar work than we would have if we didn't have gen AI just given the interest levels. And you look at all that and you go well there's a whole generation of corporate leaders in corporate America that don't know how to automate away their processes and you know, and, and, and start with yeah, that's a year long change management. No it's not. It's actually stand up the automated process and then turn off the manual process. So very different world that we're in now. I think that we get into

Chris Featherstone: this, this thinking too of you know, like for instance going to like you said, a, a non tech enterprise and their best thinkers are still within the weeds of their own cultural barriers and dynamics and everything else and legacy thinking. And so they're the ones that are quote unquote supposed to be the tip of the spear. And yet to your point, you know they're, they come in and say no, no, no, that's going to take like you said, a year to actually really get implemented. And we have to do like, if we want machine learning we have to collect all of our data at the same time and figure out how to get it into one spot. Which is not the case.

Jason Radisson: Totally. It's one of the. Yeah, yeah, I couldn't agree more. I think it's one of the biggest fallacies in IT is that everything starts with an enterprise data warehouse and a data lake and it's just nonsensical. It's very expensive, it's very self serving for database vendors. They've trained a whole generation of IT leaders to think that. But it's ridiculous. You have to work backwards from the use cases that ring the cash register and those may require you to instrument your mobile app with two or three new variables and that might be 80% of the value from your database. You need to sort of, I think always have that customer first sort of use case first mentality before you do any investments. And you know, we just see it, we see it time and time again. I mean there are entire, you know, very large cap database vendors out there that have built their whole, you know, sort of legacy on that approach. Yeah, very, very different world. I, I guess, you know, I could see there was a time where some of that stuff was relevant because you know, it just databases were so, you know, they were so expensive and so it was a big capex project if you were going to think about how you were going to collect any data. If I think back to the wireless industry 25, 30 years ago, you know, it was an enterprise project to get a little database together. But I mean these days that's the laptop, you know, not $100 million item. And you know, it's just crazy to see what companies go through and how twist up they get, you know, and often hiring in, you know, our, our, you know, new digital leader and that person comes in and spends three years building a data warehouse before they even really have any use cases that are working. Right. I mean to be, to be fair, like that's, I mean

Chris Featherstone: waterfall had its time too where it actually was relevant. But also right also is synonymous with a huge capex type of a per approach as opposed to fail fast with cloud based services or whatever that is. Right. That are super. Right, right. Yeah. And I think it's

Jason Radisson: just, it's fundamental to the startup approach and I think the reason big tech is so different from, from non tech enterprise is just, you know, they were all startups at one point. They all know, you know, there's no way, there's no way Facebook is going to forget how to do some iterative stuff to see how a product works, you know, given the DNA that they're built on. So

Seth Earley: quick question. When you start going through these enterprises and you start, you know, you've been working in both startups and large corporations, you know what, what challenges are there right now in educating executives? Because they move through part of the learning curve. But I think there's still, you know, there's still this kind of, well, AI is important. They're still not completely clear on how to, what to do with it. They're definitely not clear on what they need to do to make it work successfully. Where are you, what do you, where do you see the challenges in breaking through the noise and educating executives? Yeah, yeah, definitely. I mean

Jason Radisson: I think I kind of break those challenges down into two buckets. I think there's, there's a group of companies that has a burning platform and so AI, whether it's gen AI or whether it's some machine learning and more specific systems technology is a way to sort of fix that big, big problem or that big opportunity, get to that big opportunity that they all know about. And then I think you've got a whole class of companies where you know, they've got, they might even have existential risk around the corner. They're not really looking at it. You know, they might not have done the far out planning to kind of put those pieces together to see, you know, shoot in two years. Actually we're going to get a whole class of competitors because of this technology. They may have a lot of internal, they might have five or six items that are sort of occupying everybody's head space. And so they can't really get consensus momentum around which one to action on first. So I think you, you just, I don't know, it's, it's, it's difficult. I think you kind of have to think about what are your actionable pivot points in all of this. And you know, as a corporate leader, you know, where is, you know, where do I recognize, how do I recognize that I've got a big opportunity and like, and then how do I execute on it? And I think those are the two, right? The one is sort of the strategic reckoning, the other is like actual muscle to go spin something up. And you know, I think there's a good best practice out there which is like, it's basically the six week execution practice which is like if you have a pivot point, try to do something about it in six weeks. If you, you know, it just sort of, it's enough time to get some rough product together, you know, with five or 10 folks in a small engineering team. It's not a big corporate investment, you know, if you have to get some outside help from an engineering shop to go do that, whatever the case, but making sure that you're actually, you're doing things and you're not just talking about it and going to meetings for a year and a half, you know, vaguely keeping this stuff in mind, but rather actually doing something. Yeah, I think you hit the nail on the head when you said,

Seth Earley: you know, what's going to ring the cash register, right? What's going to improve efficiencies, what's going to have an impact and that's through understanding those processes and those use cases, saying here's our objective and then you know, what's going to enable that objective and then how do we measure it? You know, what are our baselines? Right. And how do we intervention that's going to enable that. The key though was,

Chris Featherstone: was, you know, definitely, you know, from you know, the cash back because that's the value prop. But you know, I think the critical notion of that too for, for all of us that are in this spot is working backwards from that goal and just in terms of that perspective because otherwise, you know, it'll just go on and conflate forever and ever and ever instead of, hey listen, here's exactly driving at. And we've got to be able to do that the beauty of that too is I've, we've done it in a week or we could do it in six weeks. Right. Whatever. The true goal is to see that one quick thing that's going to help with the, with the additional investment. Yeah,

Jason Radisson: right. Yeah. And that gets you, that gets you to these hard questions, the make versus buy question, the technologically possible. Right. You might come into some hardware or software constraint. Like look like the hardware that's available today doesn't actually support that use case. Let's keep our eye on the companies that are developing new stuff in that hardware dimension, you know, whatever the case might be. But it actually gets you on top of the pivot point rather than just like, like I said, kind of vaguely knowing the risk is there but not being able to execute on it. Yeah,

Seth Earley: yeah. And I want to ask you about how you've taken, I didn't, we didn't mention exactly what you're doing today. So I want you to kind of go back and talk about what you have done with other organizations in terms of managing the workforce and managing the gig economy, types of workforces and then how is that kind of brought you into this space that you're working in today and what are the lessons learned? So you worked for Uber for a while. Do you want to talk a little bit about that work? Yeah. So there's a lot, I ran a

Jason Radisson: region in the US in the early days of Uber and there's, there's a lot of, to that that's not specifically technology related. There's a lot of government relations and, and otherwise there. But what it would say, you know, specific and germane to like what I'm doing today and you know, the kinds of problems the team and I are working on. So you know, what I realized from the gig economy was, you know, if you take the software as it is, right. And essentially a marketplace for getting work done, you know, a two sided marketplace. In the case of Uber, in the case of some of the delivery apps that I've worked with invested in a three sided marketplace. You know, there's just, there's a ton of efficiency there. And so broadly speaking, what I was thinking about in founding Movo was how can we bring the kind of efficiency that we see in the gig economy to corporate America, to large foreign multinational companies that might be out there with tens of thousands of workers with a ton of inefficiency and very broadly speaking, how do we improve the lot of the frontline worker, whoever they might be, whatever inefficiency and difficulty they have in getting ahead. So I think if you look at it as a win win situation, and I always, I think of platforms and marketplaces is you got to make them win win if they're going to be sustainable and if they're going to be, you know, profitable and not just a big leaky bucket that, you know, eventually people forget about. A couple years later, you, you look at the problem set that's out there and a couple of areas of technology become really evident. And, and I'll preface it all by saying it's all machine learning. You know, I mean, it's, you look at the core of the different features and functionalities and systems. They're all machine learning systems. So basically you've got a massive scheduling capability. So one of the things these marketplaces do is they slot workers to demand for work extremely efficiently. The other thing that they do is they distribute tasks and they give you sort of a very transparent task completion, task issuance and task completion set of technologies. And I looked at those and I thought, you know, you look at the logistics industry in the US you look at clinical staff in the US healthcare system, you look at housekeeping in the US hospitality market, you look at a lot of these big job classes that employ, by the way, millions and millions of Americans and they're just, they're not digitized. And they're certainly not digitized in any kind of real time streaming, optimized way. So that was sort of the thinking that, that got us to Molo and the team and I started in 2019 with some experiments. 2020, we got very active in the beginning of the pandemic. We all had just gone full time in January of 2020. And then in March of 2020, as the pandemic really hit and the shutdowns started happening, we actually closed a deal with Amazon and started helping them with some of their most burning labor issues. And in particular in Minneapolis in, you know, everything that came with social unrest and everything that was going on in those early days of the pandemic. So a lot of learnings along the way. Along the way we've hired 700,000 people in the US and deployed them. You know, as we look at the latest iterations of our product, we're really extremely focused on the scheduling problem because I still believe that's one of the biggest drags on the economy. But that's movo in a nutshell and the kind of services that we're out there providing. And it's movo Co C O. That's right, yeah. And you're

Seth Earley: founder and CEO, which I didn't think, but it's interesting when you think about what's going on with the gig economy and you have thousands and tens of thousands of, of of workers that are out there in real time challenges. And you kind of made the anal analogy to battlefield conditions, right? Because everything is changing. You're deploying assets, you're redeploying assets, you're having problems, things are happening and then you still have edge cases that humans need to deal with. But you're taking the brunt of, of all of that brute force work of, of, of aligning, you know, supply with demand and, and lining up tasks with, with workers and, and making sure that those are completely fulfilled and, and all of the follow up and all of that stuff that happens with it. And I can see how that could obviously be applied to large workforces and at multiple levels. Right. Going beyond scheduling, you know, trying to corral and understand what all of these remote workers are doing, right. And how their work is being accomplished and how they're using automated tools. And it's kind of like this expanding universe of capabilities that you could be looking at and thinking about in terms of how do you optimize, you know, tens of thousands, hundreds of thousands, millions of workers in terms of what they're doing and how they're contributing to the organization. So it's a fascinating, fascinating space. And one of the things that we had talked about is the fact that there's an organization, there are other organizations that do this stuff, right? Like, like especially with the scheduling apps, however, one of the things, and at the core you could say function by function, feature by feature, maybe they're equivalent or whatever it is, but the big difference is that user experience, right? So when you start thinking about how do people use these things and how to. Because I, you know, we've all been in terrible applications, right? And what is it about a terrible application that makes it terrible? It doesn't align with your mental model, Right. It doesn't align with your thinking about how to solve that problem and how to go about your task. So why don't you talk a little bit about your view of user experience when it comes to the machine learning and AI? Very much so.

Jason Radisson: Broadly speaking, a lot of people are working on what I would broadly call the problem of who should work where. And you know, you can look at whether it's an ATS or a hiring platform, some of the job boards, like the indeeds of the world, you know, people are working on who fits this particular job, matching that supply and demand. What we've done at mobile is we've taken it about 10 steps further to where should this particular person with this particular skill set be working in this instance? Based on everything we know about them, based on what we know about their scheduling preferences, based on what we know about their employers demand for labor in this moment. So we integrate in real time with the labor forecast and some clients, we help them build their labor forecast and that gives us a user experience because of that real time machine learning, optimal allocation of each person's minute of work, we're in this position to give a very, very modern worker experience. So the worker, the employee, any given time has a ton of opportunity, at least the opportunity that exists in the company to go pick up an extra shift, to pick up an extra credential, to be able to work a different shift or to be able to level up. All of those opportunities are completely transparent to the worker. And they're mostly, if they're interacting, they're just interacting with a mobile app. They're not having to go, you know, the old school way of, you know, having lunch with supervisors to try to figure out who might have an opportunity. You know, supervisors don't have to go. The old school way of calling and trying to find, if you don't have the text, the mobile numbers of the different workers that are out there. All of this is just handled by bots. The bots make sure that everybody's in the right spot and the bots surface all of the opportunities that are out there in the company. So it's a long slog to sort of get all the machine learning in place to be able to do that. But once you have that kind of capability, it's such a modern, light, easy experience for the employee that nobody wants to go back. And so, you know, I look at it as like, in some ways, solving scheduling is one of the best things you can do for worker engagement, for giving them the feel of feeling of empowerment. They're more empowered over their own time and flexibility. They're more empowered over their career path than almost anything else you could do. I'd love to get you to build workforce

Chris Featherstone: management systems for call centers. And I've always loved the gig economy perspective, right, of this, like, hey, I want to self choose. What have you seen though, is the inverse impact of that on the psychology of almost borderline entitlement empowerment? Does that make sense when we're thinking about those? Especially with an organization that's coming from years, like decades or Centuries of culture that is very much steeped in if work for 40 years or whatever. I'd love to get a take for what you've seen in psychology change in that. Oh, totally. And you know, and

Jason Radisson: I'm not sure how many HR leaders might be in your audience as well. I think it's very relevant for them. Also. Look, I think there, there are guardrails. Most people focus on the guardrails. The guardrail on the one hand is the employer guardrail and it says, this is a schedule, take it or leave it. I think the other guardrail is sort of the, you know, completely gigafied. Hey, this is when I'm going to work this week. Take it or leave it. Employee perspective. The way we solve it with most clients is we will fill the schedule based on your preferences. However, if we don't have anybody that wants to work a shift, we're going to put somebody in it. And so, you know, it's somewhere in the middle. We're meeting employees somewhere in the middle. And like I said, I think success long term of any workforce, really any, any marketplace, but any workforce is, it's got to be win, win. And you have to kind of find these, these kinds of solutions. And I think, I think we've got it now. That doesn't mean that we don't have employers that see it a little bit more on their side. And you know, we're still figuring it out on a case by case basis, but it's basically that it's. And it starts with the employee experience, starts with the employee putting their preferences in that then mostly get honored. The other thing too, we automatically do a bunch of substitutions. We do substitutions, we do peer to peer shift swapping. So even if you don't really get your preferences in your main shift, your main shift is 9 to 5, Monday to Friday. But you would really like to have Thursdays off. But the company says, too bad, sorry, we can't honor that because we need you on Thursday. Right. You can still put in to be an alternate on a bunch of other shifts. Say you want to pick up some weekend hours and you get those. So even if we can't always honor 100% your preferences, we might be able to open up some new space and some new earning potential that you wouldn't have been able to get otherwise. So, you know, it's not just black and white, like, you know, no on Thursday, but more case by case. And I think that's just life. And I think the best supervisors have kind of always Done that. We just, in creating the system in the way we have, we've been able to. To replicate that in, in this automated way. On Thursdays. We

Seth Earley: won't say no. We'll try. Right. All right. And maybe one Thursday,

Jason Radisson: slow. So, you know, maybe you get a Thursday a month off. You don't get all of them off. So to us, that feels modern. And I think it, you know, we've been talking. I mean, it's, it's a, it's a compromise. I think we're all sort of dealing with generational shift, right? No, you know, we go to X, to Z, to X, Y and Z, and people have different expectations. And look, it is not going to be 100% the employee guardrail. Nobody can run a business that way either. And of course, it just is what it is. I think at this point we found. Too, some of the. We'll say

Chris Featherstone: older generations were able to leapfrog into a gig economy thinking because they had been out of the workforce for so long that they kind of skipped that generation of, oh, they used to be super, you know, like cultural 9 to 5, and then they're out of the workforce for a while and now they jump back in. And now it's this perspective of complete flexibility in theory. But, yeah, it's been really interesting. So, yeah, that was the hardest binary problem to solve in a lot of cases was here's the plan. However, when the plan hits reality, you know, and you have people that are out sick, and then, you know, Corona, you know, provided a whole new perspective. Right. Just the. The binary goes out the door quick around like, oh, it has to be flexibility in there, you know, very much.

Jason Radisson: I think there's some really interesting. And as you guys know, I've worked in China and in particular with Didi, who bought my last company. I've also worked a lot in Latin America, and I think there's a lot of interesting stuff coming out of the developing world. Large companies in Latin America see the gig economy as the competition for labor. They go, if you're running a bottling plant, you're a massive bottling plant. You're competing with Rappi and Uber for employees. And, you know, all the employees that are coming in, all your front line is used to having everything be based in a mobile app and having that complete flexibility. So, you know, I like to think that with the solutions that we're finding with each employer and sort of where that slider is, that we're actually providing an even better experience because, you know, the flip side of the gig economy is you have perfect flexibility to pick whatever hours you want. But you may pick really crappy hours, you may pick an hour where there's, or a schedule where there's no riders or there's no orders to be delivered. And so you make no money, you know, versus, you know, if a hospital is having you come in, it's because they have work for you. It's not like you're not going to get paid in our world. So that's awesome. I have a question about

Seth Earley: organizations taking risks and looking at where they are in terms of their AI journey. And one of the things we talked about in our prep call was that, you know, you'll talk to some executives and we'll say, well, you were getting our infrastructure together and our tech stack and we're doing these things and we're not ready to do that. And, and I hear that a lot, you know, we're not ready, you know, we're, we're still trying to figure out, you know, what the, what the bet is. And so you had mentioned the fact that there can be million dollar bets and there can be billion dollar bets, and I think you meant it in terms of the implication for the organization. And even if we say, even if we bring it down in order of magnitude or 2 and say, you know, $100,000 bets versus million or 10 million, you need a couple of different things in order to do that. Number one, you need an executive that believes in it and that has enough social capital to be able to affect change. Obviously you need budget for that, you need the right set of use cases, but you really do need that kind of expansive thinking about where the organization can be and what it can do. And do you want to talk a little bit about that mindset and what you've experienced? And I think it was really interesting the way you said, you know, multi million versus multi billion or even again if you say 100,000 versus million. But talk a little bit about that mindset and how organizations need to cultivate that and executives, how executives need to go into these things, thinking about potential impact. Totally. Well,

Jason Radisson: you know, to try to put it a little bit as a, in my answer, a little bit as a pivot point. I think it's, the pivot point is if you're in a company and there is a lack of a common understanding of what the big automation opportunities are. So my challenge to anybody in a large corporate environment would be, can you name off the top of your head four or five big automation opportunities and put dollar Opportunity like dollars behind them. Like we were saying, this is a $2 billion opportunity, this is a $20 million opportunity, this is a $200,000 opportunity. Because I think that's the, that's the place the mind, sort of the mindset, the headspace you have to be in, and then everything just becomes execution. And, you know, in terms of change management, yes, you're going to have to, you know, figure out the path that your company and culture and everything demand to sort of get there. But at least you're clear on the what. I think it usually starts with, with the lack of even understanding the what. One of the biggest examples that takeaways and examples from my career. So I worked for Gary Loveman at Harrah's Entertainment and then Caesar's Entertainment. I was the post merger guy. I ran database marketing there and the statistics and algorithms team and post merger management. And when we bought Caesars, part of the justification would involve publicly traded companies. We had literally twice the EBITDA per square foot in our company that they did. And the reason we had it was, was largely due to an algorithmic approach to base management. Like we would get a new customer, someone would come into our casino for five seconds and we would figure out what they were worth, and then we would start disproportionately spending on them, on the customers who were worth it, who had the upside. And we did that just famously. There's a bunch of HBS case studies out there and our business review articles and other business, press and academic work that's been done around that. But the crazy thing was it was all published. Literally our algorithms were published for years and years. And literally Caesar's Entertainment, right across the strip from us, didn't employ any of the strategies. I, you know,

Seth Earley: that and boggling. Go ahead. And that's, that's sort

Jason Radisson: of, that stuck with me my entire career where I say one of the biggest problems in corporate America is a lack of sort of understanding where the big levers are. And, and it's not an individual thing. It's not like there's nobody at Caesars that ever understood that. But it was, it was, you know, sort of an enterprise level thing that the company couldn't sort of click the right pieces into place in order to execute on the playbooks that we're winning in the industry. So, so this is where I

Seth Earley: famously say, at least famously to myself, that having the answer doesn't matter, right? Like I said, right, Having the answer doesn't matter. Like I, I can publish articles and Say this is the way it is and here's how you do it and here's the case study. And yet there's this, you know, there's this weird like gap. And, and I'm so glad to hear you tell me this. Right. Because it makes me feel like it's not just me that, that you had published these things. They showed the value, HBR picked it up, all of these. I mean I would want to go back and interview those executives and say what were you thinking? Or what were you like, why, why did this, did you not know about this? Like, had you heard about this? You know, I mean, I'm sure they have. Right, right. Has anybody done any kind of forensic discovery or research? I mean, you know, it's. Yeah, yeah. It's

Jason Radisson: so totally, totally, so long ago. But I can, I can talk a little bit about, you know, sort of the experience in the post merger, you know, clean room but then, or non clean room but then sort of the post merger stuff. But what I would say, take the. Example when you say clean

Seth Earley: room in that context you're meaning regarding data or what do you, how are you using clean room? What I. Is in the pre,

Jason Radisson: pre merger phase where the two companies plan together. Right, right. Yeah. And, and so, you know, we got a good look into things. And I'm not, you know, like this is like the companies have changed owners numerous times and, and there have been other books written about sort of, this is all public information, everything that's happened. Yeah, but, but let's take the example, an algorithmic example. Let's take, let's take yield management. So the pricing of rooms, you know, so there's. For those that are out in the hospitality world, there's an old school way of doing it which is there's literally a guy or a gal who prices rooms. Right. And it's the, it's the mom and pop way. And it's like here's our calendar and here's what our rooms cost. And that was replaced kind of famously kind of with some technology that came out of commodities trading back in the 80s and with American Airlines and with like the real time pricing of airline seats. And then the casino industry deployed that through. There was this company called Probes that was running around sort of selling real time room pricing. And at Harrah's Entertainment we had nine segments and we priced our rooms in real time. And we had kind of done this and this was sort of a top down decision, you know, sort of since I think Gary Loveman sort of came to the company. If you looked at Caesars Entertainment, they had manual sort of old school room pricing pretty much everywhere in the company. And they were working on an extremely complex 99 segment room pricing system that hadn't really been deployed and 99 cent. The problem with 99 segments is you don't have enough traffic. You just get into kind of very messy answers with very low statistical significance. So we'd long sort of rejected sort of a extremely fine brain model. Nine segments was enough for us. But I mean, we can talk about, you know, other things that they kind of got wrong. But like fundamentally like they, they sort of, they sort of made and you know, like I said, this is all these people, all the roles have changed, Everything's changed. This is, you know, 20 years ago, 20 plus years ago. But I think the important learning is like they made a, essentially sort of a extremely half hearted effort to even sort of pay lip service to what needed to have happen. And you know, that I think is where a lot of this stuff, you know, gets bogged down. It's like, you know, even if the strategy, even if there's pressure to have the strategy in place, you know, have you read in what should they have done? They should have taken our model and exactly replicated it in one casino and proven the difference, proven that they could get the 2x lift in revenues per room and EBITDA per room. And then, you know, had the momentum to be able to have done more change. It might have been a very different sort of history of the industry if they had done even that simple thing. So I, you know, it's a long story, but the moral is really short, which is you've got to, I think you've got, you've got to know the what. But to your point, lost knowing the. What isn't enough audio there. Jason, can you hear?

Jason Radisson: Yeah, yeah.

Okay. Huh? Yeah. No, it sounds like I've lost audio.

Chris Featherstone: You know, I, I think. Yeah, I have lost your audio. Here we go.

Seth Earley: You're back. Okay. I don't know. Yeah, yeah.

Chris Featherstone: As, as they say, you know, from the old, you know, from what you were saying before, you miss 100 of the things you're not looking for. And I think that's what happens is you get, you know, these, these organizations get in the weeds and can't figure out, you know, what the hell to do because they don't know what to look for it. Right. And to, you know, to your point, they don't have anybody that's there to say, hey, by the way, did you see what they're doing across the street and is that interesting? Should we think about that? And usually that's lower, you know, below the budget line type of folks that are seeing that kind of stuff. But above the budget line, people think they've got it made in the shade, so they're doing it and they just.

Seth Earley: We're getting to the end here. So, Jason, tell me a little bit more about your world. You know, do you have kids? Do you have hobbies? Do you have pets? Do you travel? Do you. What else do you do besides work?

Jason Radisson: Yeah, family and outdoors are really the things that I do outside of work. I have, I have four kids from 18 to three. We're busy on that front and we moved to Minnesota during the pandemic and, you know, I've kind of grown up in the outdoors and do everything from skiing to fishing and hunting and I'm just out in the woods with the kids as much as I possibly can be outside of work. So after the three year old, what's the next

Seth Earley: age? What's the next.

Jason Radisson: Yeah, yeah. People often think they're twins. Yeah, yeah, yeah, exactly. We re up. Yeah, yeah.

Seth Earley: This has been great. I've, I've really enjoyed chatting and what do you see as the next frontier? You know, when you talk about workforce management, like I could see this being the beginning of a lot of stuff. Like in terms of optimizing a workforce, what do you see in the horizon? What's the future look like? Well, I mean,

Jason Radisson: I think it's funny to watch the evolution of Gen AI and I think you can't help but look at Gen AI and realize that it's a general technology that sort of needs to come back around to solving real specific use cases. And I think that evolution is going to happen. We saw it with machine learning. I think machine learning was a little bit stealthier, a little more under the radar and just kind of done. You know, there was just a day when we all woke up and Amazon was there with a real time optimized homepage and we kind of didn't realize it. We didn't. There was just a day we woke up and American Airline was charging us a specific price for a seat. There are so many fully automated, very smart systems in the world. And I think what's going to happen with Gen AI is it's just going to come back around and it'll enhance some of those engines, it'll sort of support a ton of new use cases. But I think that process is just in its infancy. I Look at what that means for work. It probably means some of the same things that we've seen in machine learning companies, which is we need a lot more of these sort of operations research, industrial engineering job classes. I've been bullish on that, you know, university program, college program for 20 something years. You know, I used to speaking of Caesar's Harrah's lead campus recruiting for undergrad and for business school and grad school and those types of roles are as hot today as they ever were back then. I think those are the people that are going to be leading these systems and that's where a lot of the job growth is going to be. And I'm very bullish on the frontline workforce and I don't think we're going to be replacing nurses, home health care people, a lot of hospitality service people. Any of that is going anywhere anytime soon. So you look at all of these advancements, probably the role of supervising and you know, the we're going to see more sort of remote plus forward deployed resources I think is just this new evolving model of work. And you know, this is the new reality. We are, you know, half the time going to be talking to people and running teams on Zoom and you know, and other tools like movo and you know, the there's going to be a little sliver of the old world but I think we're already pretty much into this new mode and I'm just excited for everything that it needs. It's

Seth Earley: going to change quickly and it's mov co and you can connect with Jason Dash Radisson R A D I S S O N that will be in the show notes. I really want to thank you for joining us today. Jason, thank you so much for your time and your expertise. Thank you guys, Seth, Chris,

Jason Radisson: both for having me on the show. It's been a real pleasure chatting with you.

Seth Earley: Great. Thank you to the audience for listening and tuning in. This has been another episode of the early AI podcast and thank you for Carolyn for her behind the scenes work and Chris and again thank you Jason. We'll look forward to continuing our conversations and collaborations. So really enjoyed it today. Thank you.