AI for Effective Workforce Management

It is a pleasure to be here with you today, and I'm very excited about today's discussion. Uh, Liam and I have been talking about all things work, especially remote work and people analytics for a while now. And it's, I I get excited every time we speak. And today's discussion is really focused on that AI for effective workforce management. The transformative power of AI has already had a dramatic impact on the modern workforce management and will continue to in the future. Our guest today, Liam Martin, is the co-founder and Chief Innovation Officer of Time Doctor. Uh, and he will be explaining the potential for AI and people analytics, predictive analytics, and smart decision making. And this all, all really connects and especially when it comes to quit rates and retention and, and really making sure employees are engaged and successful at work. Liam is the founder of Time Doctor and it's an innovative enterprise software company specializing in productivity enhancement and time tracking features that are highly acclaimed for their employee friendly capabilities. He's also the author, just like I'm an author of a great book Running Remote Master, the Lessons from the World's most successful Remote Work pioneers. He has also has a conference called Running Remote with some unbelievable guests like Nick Bloom, who has tons of research as well when it comes to remote working. So, Liam, it is great to have you here And thank you. Thank you very much for having me. I am also really excited every time that we talk, so I think this will be a really interesting conversation, uh, for everyone involved. Again, get your water. I have a big water bottle here. Um, get your notes out because this is stuff that I have not released ever in the history of the last 15 years that I've been working in this particular industry. I also just saw someone from Trinidad and Tobago. Is that true? If true, if true, that's amazing. 'cause that's where, um, that's where my family's from. Ironically, I'm the last white Trinidadian to be in Trinidad and Tobago. Oh, and by the way, make sure you post your questions in the q and a so that we can make sure we get to them throughout this conversation, and definitely towards the end. Thank you. Yeah. Yeah. Thank you very much. Um, so we're gonna get into AI for workforce management. And me and Dan talked about this quite a bit, which was, we can definitely talk about where it's going as an industry, but I wanted to actually very much focus in on some really clear outcomes that I think are happening inside of the entire HR space. Because to be completely honest with you, I, I think we're going through one of the most times in the history of all HR as an industry. And I think that where we're going to be in the next five years is going to look nowhere like where we were five years ago. I don't want to scare a lot of you. I want to be able to warn you or give you, uh, be your advisor in this context because AI is not going to replace you, but someone using AI will be replacing you and you need to be able to understand that moving forward in terms of everything that we're currently doing. So, as Dan said, uh, this is me. This is the standard slide. Please don't leave this session. I know what I'm talking about. I'm the co-founder and chief Innovation Officer at Time Doctor, which is a time analytics tool for primarily remote workers, but we manage people that are in a hybrid relationship and in office environments as well. I'm also the co-organizer of Running Remote, which is the largest conference on building and scaling remote teams. We've been running that for six years before it was cool, by the way, before everyone jumped on the remote work bag bandwagon. We were there. And then I also am the co-writer of a Wall Street Journal bestseller Running Remote, which looked at the methodologies that led to successful remote companies before the Pandemic. So what did all of those companies that were working remotely before the pandemic had in common? It was one singular thing. You're gonna have to buy the book to find out what that is, however, 'cause we're not talking about that today, uh, but I might actually be handing out a couple of these at the very end of this presentation. If, uh, you stay till the end or maybe not, you'll have to find out. So I'm really focused on understanding how work works. That's what I've committed the last 15 years of my life to. Before that, I was actually a academic, um, although a very bad one that was focused on the sociology of work. So for me, this is easily the thing that I'm very passionate about, the thing that I'm up at three o'clock in the morning about. And also the thing that you can ask me a lot of questions about. So led me to writing this book, uh, which people like Cal Newport said, is a critical and much me much needed guide to thriving in the new world of work. We have people like Matt Pnis, Scott Young, uh, we have people like Kim Scott have all Sarah Sutton who have all written, uh, reviews of what value this particular book has brought to them. And I'm happy to be able to talk about that afterwards as well, if you're interested. We have team members in 48 different countries all over the world. Uh, we work remotely in a very specific way that has allowed us to be able to do that. And a major part of how we've been able to do that is through implementing AI in terms of workforce management. And I think the question that I was asking at the very beginning of, or the top of the hour, is really important to be able to reinforce again, how does our industry not just survive, but thrive in the age of ai? As I said before, you're not going to get someone, you're not gonna get replaced by ai, but someone that is using AI could very possibly be replacing you. So how do we actually make this transitionary period? And I think we can actually answer that very simply through predictive workday analytics, which is where I really see the evolution of our entire industry going. So let me give you a little bit more of context on that. I don't know about you, but three years ago there was like no remote work. Essentially it was about four and a half percent of the US workforce was working remotely pre covid. Uh, there was no AI in si in sight, so we didn't have chat GPT, we didn't have everyone and their mother building AI companies to be able to disrupt the industry. And then we also didn't have a economic, um, problem that is always seeming to be one quarter away, but is never really solved. So fundamentally for me, like no other time in history, I think that work has changed and, and really it's also the perfect storm, right? It's all of these absolutely kind of trends, you know, economically, socially, technologically that have come together that have enabled this transition to happen. Absolutely. Uh, great point. And I know probably me and Dan think about this quite a bit, but as someone who follows technological trends in our industry, I felt a little bit like this guy actually. And, uh, this is something that I put together in about 10 seconds on an application called Mid Journey, which allows you to be able to build any image you want instantaneously. But more specifically, since I work from home, um, I'm more like this guy. So this was built by Dolly in about 35 seconds, seconds. And very specifically, this is what I told it to tell me. Uh, scared, comical guy working from home with a headset on trying to do his job, overlooked by the four horsemen of the apocalypse. And funnily enough, he actually kind of looks like me. And they, the AI didn't even know that I worked for Time Doctor, but it put time doctor down as the logo because I've asked so many previous questions about time, doctor in the past, this is the worst that this technology is going to be at. And this is a very, in, this is a very scary to some perspective, to me. It's incredibly exciting. But this is the scare. This is the worst version of this that currently exists. Think about this in a year, think about this. In five years, this is gonna completely transform our entire space. So what is predictive work analytics? Well, very simplistically, here's a perspective of how I think it kind of works its way out. Oh, I love this part. I've got no address last known or otherwise, no tax returns for the last five years. Check NCIC, maybe he's got a record and senate protection team. As soon as we lock location, it looks like federal housing, concrete glass, egg crates. Ouch. But a thousand of rows in the district Fractured him just coming in. Numbers nine. Okay, so for anyone that didn't know, that's Minority Report, and it's funnily enough, funnily enough, someone literally a month and a half ago told me it was like, oh, I feel like I minority. I'm in, I'm Tom Cruise in Minority Report. Uh, that's literally what predictive analytics is like, how can you predict what people are going to do before they do it? So it's empowering companies to measure knowledge workers output to predict outcomes and increase productivity or decrease negative output inside of the organization regardless of whether you're in an office distributed or you are hybrid. Essentially what you would have is something like a hypothesis that work-related stress can lead to burnout and eventual quitting of your job. An indicator might be working late on emails, maybe a sign of too high, of a workload and stress buildup. And then the variables that you would look at would be total time tracked on email or application or websites after 6:00 PM in the user's time zone over a 30 day period. And then you would check whether or not that person has stayed or quit their job as an example where they have higher elevated levels of stress. Now, I'm actually a little bit concerned about putting this kind of stuff out because this has never been seen before. Uh, outside of some of the internal stuff that we've done inside of our company, there are other companies that are kind of working on this, but no one I think is as far ahead as we currently are. So we use a lot of, um, decision tree analytics. We actually use something called Bayesian algorithms to be able to analyze all of the different reasons why people would quit their jobs, as an example. Or another variable might be whether or not someone is successfully onboarded their job or whether or not someone's ready to be a manager. So we look at all of these different variables, and there are tens of thousands of these variables that impact this model. Uh, just for quick prediction as an example, it's like if you've worked more than 251 days, uh, inside of an organization, your chance of quitting goes down by 10.2% based off of our machine learning algorithms, just as one clear example of something that we do. And so from that, we can actually build an entire algorithm that tells you specifically when someone will quit their job. And I'm using this as a small example, I don't wanna tell you, like we can't go through all the other variables, obviously, because they're a little bit more complicated. But things like company culture as an example, are you a new employee? Um, are you working late in the evenings? Are you working early in the mornings? Are you working outside of your regular work hours? That's actually like 25% of an entire quit model identifying whether or not someone is going to stay in the organization or quit and find other work somewhere else. Are they working on weekends? Are they overworked? Uh, do they have lack of consistent work hours? There's, as I said, thousands of different variables, but these are some of the ones that directly that most impact the overall chance of someone quitting their job or not quitting their particular job. So do you see, do you see any differences here between companies of different sizes, companies that are hybrid versus fully remote company? Uh, yeah, individuals like who are, you know, different, uh, you know, diversity types, uh, yeah, so gender and, and ethnic, All that. So funnily enough, these algorithms are actually completely rebuilt, dependent upon the size of the organization. So an algorithm for 25,000 seat plus organizations is actually completely different than 250 seat organizations. So there's like a really interesting data differences that we see between those. The biggest ones being, let's say just between a 25,000 seater and a 250 seater is bureaucracy. So there's a lot of middle managers and we see a huge spike in essentially u uh, employees being p****d off at their managers more in large organizations than in smaller organizations because they're able to keep that culture tighter fit. And that actually kind of connects into user and company traits, which is essentially company culture. So if you're an HR professional as an example, it's so critical because this is a number that you can actively move inside of a company, and as you get bigger, the problem gets bigger. And so being able to actively manage that is, So it's almost like if you, if you shared this with a smaller company, as they're growing, they could use this almost as a guide to say, Hey, if we focus on these key areas, then we won't have as many growing pains per se. Exactly. Yeah. So we've had, um, I'll show you an example actually that just very clearly alludes to that, and we couldn't get any of the data from some of our clients, but we actually have some of our own data, uh, which I'm very excited to be able to share with people. So here are the six most counterintuitive insights with predictive analytics. And for people that answer these questions successfully, uh, the person that answers the most questions successfully will definitely get a copy of the book. And maybe someone else that answers really interesting, uh, gives me really interesting answers. Would also get a copy of the book too, Dan, you can check it in the comments to make sure, uh, that everyone is, uh, playing along. So first one, here's one that we didn't really understand. Quits spread like a virus. And for some HR professionals, maybe this is just super simple, but for us it was a really interesting insight. So if someone quits inside of the organization, there is a 58% higher chance that someone else will quit inside of the organization. So quits spread like a virus, and we haven't really been able to quantify this before, but in our algorithm it's 58%. So if you have 10 quits, then the chance that 10 more quits will happen is a, has a 58% chance, 58% higher chance than if those 10 people didn't quit in the first place. So it's almost like a snowball. And what you need to do as HR professionals is to be able to stop that process, you need to be able to say, Hey, let's address this as quickly as possible. I mean, this is kind of HR 1 0 1. If you're gonna let somebody go, if you're gonna let someone go, you need to be able to just make sure that you, you let go of that person as quickly and as efficiently as possible without disrupting other team members. But no one really thinks about the quits in that same context as well. So this model shows that very clearly actually, if someone wants to leave the company, giving them let's say two months to be able to move to the next, uh, to the next to their next stage in their career, is probably not the right decision to make in reality. Actually, you might wanna cut that down as tight as you possibly can. Like do one week, do a couple days if you possibly can, because we've actually identified that the faster that those quits leave the organization, the lower this chance of quits, uh, current. Do, do you Think it's like a few, a few, a few things with this is like, do you think it's people feeling like, oh, like I could be next, so I should almost quit to save myself from getting, from getting fired. I should start looking for my next job. I see too my, you know, coworkers leaving. Or do you also think it's all, it's like, you know, this certain amount of work that needs to get done, and as people are quitting, there's fewer people doing that same amount of work. So burnout, like you were saying before, kind of shoots up and then those people just end up leaving as well. Yeah, The unfortunate thing about an algorithm is those things could all be true, but they require a lot of qualitative research to be able to actually collect those types of responses. But at least now we have a pointing indicator to be able to identify like quits right after, let's say someone in your department quits and then two other people quit. I would actually pay very special attention to other people in that department right after that termination or that person leaving to be able to make sure that the rest of that department is doing well, just in terms of HR touching base with them, trying to see if they're frustrated, if they're overworked, if they're burnt out, those types of things. So you're absolutely right. Yeah. Yeah, go ahead. Here's a good comment. Also, people may feel that the person quitting knows something they don't know and then another person that all of the above plus, uh, summit gives them the permission and courage to move forward. Yes, and I think that that's another piece. Again, it's awesome to have a whole bunch of HR professionals in here to kind of like unpack what we we see from a quantitative perspective to get the human component into it. But all of those things very well might be true. And I think you just need to be able to go in and do that analysis yourself to be able to figure it out. Next one here, uh, overwork does matter, but not for the reasons that you would actually think. So this is a really interesting one. Our algorithm identified that overwork was the third most important overall variable leading to quit, but specifically a one-time, a one-time increase, uh, a one-time increase to their chance of quitting more than all their sources of their overwork combined was just one single variable, one single variable connected to re remote to overworking produced almost all of the quits that we saw. I would love if anyone could put it down on the comments, what is the one single overworking aspect that leads to the vast majority of quits, at least in our algorithm, and we have a couple hundred thousand quit events. So we, we've got a pretty clear understanding of what this is. What do you think the one single variable of overwork was? Oh, there's so many good comments. Quitting issues with manager, lack of acknowledgement. Uh, a lot of people say manager, that that would be what I would say as well working over holidays, because you know, there's a saying, you know, you don't leave a company, you leave a manager. Yeah, yeah, close. Working on Working on nights, lower recognition or lower rewards. Uh, limited flexibility, no recognition. Again, a lot of no recognition. Okay. A lot of no recognition and appreciation. Wow. Yeah. Restructuring. None of those. None of those. It's great. Lack of Purpose. Okay, this is great. None of those, here it Is. No extra money. Poor workforce planning burnout. Well, I mean, connected to it, but yeah. Okay, here it is. Got the feedback. No one got it. So no one gets the book. We'll have to wait till the next question. Friday nights. So overworking on Friday night, if you work late on Friday night, that actually re increases the chance that you're gonna quit by 52%. Working Friday nights not work. Working on weekends does increase your chance of quitting, but not anywhere close. Like we're talking like one to two percentage points. Friday night is by far the worst thing that you could possibly do to your staff to be able to force them to quit due to burnout or any other case, right? Just to be basically create a quit. So this kinda was like huge on our model, and I actually have a human story connected specifically to this. So this is our algorithm as it applies to quits inside of time, doctor. So you have like lack of consistent work hours working in the evening, working early in the morning, uh, work on weekends, long frequent breaks, short work days, all this kind of stuff. And we had this employee, David, and so David, uh, this is his chance of quitting on any particular day. And you can see that 30% is okay. The standard deviation of someone wanting to quit is between 25 to 40% on any singular day. So people are always throwing off indicators. If you're in the 80% range, this is a time when you need to actually like, have a conversation with that person and figure out what the heck is going on. So David, um, around November the fifth, one of his coworkers quit and then David ended up getting extra workload and David was not very happy about this and David actually, uh, told his manager, Hey, listen, I'm overloaded. I'm working on Friday night. And then look at this work late in the evening. This is David's data. This is dated David's algorithm. So this is our entire company algorithm, and this is David's reasons for wanting to quit, lack of consistent work hours, working on weekends, th weekdays compared to the control. Uh, working early in the morning and working late in the evening. He was working, working on Friday nights and he did not like it. So this message was sent between our CPO and, uh, his, uh, their middle manager saying, Hey, David's got a big problem, coworker quit. He's working nights, he's super stressed out. So that's on November 7th, again, we're here, we're at November 7th. We sent him a letter, or sorry, we sent him an email November 13th saying, David totally understand what's going on. Number one, after 5:00 PM we're not allowing you to work. Like, we're literally gonna cut you outta the system at 5:00 PM Um, we're gonna have someone work to get a replacement for Q four, latest Q one. In the meantime, we're gonna have another employee come in and help you to be able to work on this. And then also you need to be able to track all your data so that we're just very clearly identifying your work hours and identifying whether or not you are, um, still need, still need more help. So he sends an email like two hours later saying, thank you very much. Really appreciate it. And here's the result. He went from 46 just from that email went from 46 to 23% in terms of his chance of quitting. So we saved that employee from basically quitting or at least being very frustrated and, uh, and wanting to leave their job at 52%. It's basically a coin tos whether or not someone will quit their job within a six month period. So for us to be able to get them down within that was incredibly powerful and just showed that sending even a message, very few HR professionals have the quantifiable evidence to be able to lead to their sending one message equaling impact in terms of their employees engagement. And we just thought it was really cool. We have a lot of other stories for our clients, but we can't use theirs, so we can only use our data. Well, and The data's useful too because it's, it's sort of non-biased in a sense where you could show this as an HR person, a leader or professional, and say, Hey, absolutely. Like this is the cause of it, let's stop doing this. And it really is a transformative thing for a corporate culture. Right? And can, yes, you said, you know, obviously, you know, in a bad way it could be more of a virus, you know, burnout and, and people being unhappy and disengaged, but also in a positive way it could spread good feelings if, you know, there's a culture that supports employees who might struggle with burnout, right? Absolutely. Yeah. So the big Question I have is, do you layer, have you ever laid on pay? So like we did a massive global study and we found that mm-hmm. The two big things that are causing burnout is one is overwork, like working for instance on Friday nights, but also mm-hmm. Pay is another component. So if, if someone's getting paid more, are they more willing to work on Friday night and and less likely to quit? But if they're not getting paid more, if they're getting paid, you know, I guess industry average or below average, are they much more likely to quit because it's harder for them psychologically to justify the additional work on a Friday night? Right? Yeah, no, it's, it, that's such an interesting perspective as well. We haven't pulled in, uh, pay tiers because we just don't have access to that data. But it is actually one of the plans that we want to be able to add in your workday analytics, your Bamboo hr, whatever it might be to be able to add on to this, to be able to refine this algorithm. As I said before, we have a 74.2% chance of predicting, successfully predicting someone quitting their job within a six month period. But that's untrained. And when we train the model specifically on that organization, uh, the top one that we have right now is 90 plus. So we can get this really tied in just to your org, and we need about 200 quits to be able to actually process out that data. Basically the algorithm needs to figure out what are unique, what are the unique differences in your company in comparison to all other companies that we take a look at. But, uh, it's still, and obviously adding in pay will definitely boost this up. And, and Jennifer says, she goes, I work for an org that made everyone leave at 4:00 PM on Fridays. Even when we had a seven day a week month turnover was low. So she basically is backing up this data point. Okay, that's cool. Yeah, I mean, and it's just shows the algorithm that it's, it's so fun to be able to look at this 'cause like this quantifies hrs activity. HR is always one of those things that's that's a nice to have. Yeah. Okay. Like people are not quitting. We're hiring people. Okay, cool. Like how much did you actually add into saving this particular worker's, you know, job or them leaving or having to replace someone else. It's really cool to be able to have quantifiable evidence that essentially an email produced a 50% drop in someone's chances of quitting. Uh, which we're really happy that David is staying with us because he's really, he's really important to the company. Uh, we can get into that later. Next one. Here's one that, uh, I think is kind of, I mean, it makes sense, but not necessarily in the way that you would think. Boredom leads to quits a lot. So our, a AI identified that essentially employees being bored was a significant factor towards them quitting. And I have some variables connected to whether or not that the top variables connected to people quitting due to boredom. What activities do you think lead to a increased chance of quitting due to boredom? I'd love to see that in the chat. And I, I'll give it like 30 seconds for people. And again, whoever uh, gets the right one will get a free copy of the book. We have not enough responsibilities, tasks lack. Okay, wow. It's gonna move the best. Lack of variety. Unnecessary manual process, useless meetings, too much repetition, online shopping, repetitive tasks, repetitive tasks, uh, lack of career growth, overly structured processes, not enough creative freedom, busy work, lack of challenge. That's what I was gonna say is lack of challenge. Okay. Meetings, uh, not allowed to try new things and fail, uh, unclear expectations. Not challenged, not challenged, not feeling challenged. A lot of not feeling challenged. I would say that's the number one most popular response. Okay. Okay. All right. Lack of engage. Yeah, when You said boredom leads to quits, I immediately like, oh, like that is a way people get disengaged in their work as well. Yeah. And, and it's, it's very specific things that we can now detect now with at, at least as it applies to this algorithm. So here are a few higher level of time spent in your calendar app, huge impact whether or not someone is actually gonna quit. Now, why would that impact quitting? Well, I think it's because they're spending more time figuring out what fricking meeting am I gonna be doing next and how do I have to deal with all of these meetings and I don't wanna do these meetings anymore and I'm gonna quit, right? So, and a significant increase in the amount of time that they spend in their calendar results in them quitting, um, lower levels of interaction on their top three most used applications. So that's generally something like their email, their instant messaging tool, and then something that is connected to their profession or their industry like Adobe for graphics designers as an example. If those three drop, that's a significant indicator that someone is gonna quit their job. Um, increase in breaks and break time is another big variable. And then more time spent on most used applications, but less activity is the biggest indicator of increased predict. Quit prediction. So you spend more time on your email, but you actually answer less emails while you're on email. That is the biggest indicator that you're going to quit 'cause you're just bored. 'cause it's like, uh, well, you know, I'm gonna answer these emails or not answer these emails. I don't really care 'cause I don't like my manager. Or, you know, I'm not feeling challenged or I'm just stuck in meetings all day long. So those were the top ones that we found based off the algorithm. And, uh, Did attendance relationship, did attendance come up? Did attendance come up as in like people instead of starting the workday at eight, they start at eight 30 or or 15 minutes late or a half hour late because they're like, So it's coming up. Do it. Yeah, That's the next one. Uh, yeah, that's, that's, that's another one that's coming up. All right. So measuring engagement is only half the battle. All right? Engagement outside of top websites and applications is by far the largest indicator of somebody wanting to quit. But measuring anything off the top three apps has no impact towards them quitting, which I also thought was quite interesting. So for most people, their most use applications are email their instant messaging tool and as I said, what they do in their profession. So it's like Adobe or their CRM if they're a salesperson, or maybe for me and you Dan, it would be Zoom, right? 'cause we're always in these Zoom calls, uh, left, right, and center, right? So it's like less activity on those three is very important. Anything outside of that has almost no statistical correlation towards whether or not you're gonna quit your job, which I thought was also really fricking cool. Um, next thing here. Some overwork is actually good for reducing quick chances. I know that that sounds very counterintuitive and probably a lot of HR professionals will shutter at that, but the data is very clear. So we found people who had consistent work hours had the lowest chance of quitting. IE variation is the biggest impact on quitting, not how long you work. So if you work like four hours one day and then you work 11 hours the next day and then you work six hours the next day and eight hours the day after that, that is really bad towards someone having a higher chance of quitting their job as opposed to everyone, uh, one person working six hours every single day. That's more consistent. Whereas inconsistency leads to a lot more quits. And let me give you a couple examples. Biggest variable decreasing quit chance is working on email before 9:00 AM. So if you're working on email before 9:00 AM you're great. Like you are never gonna leave this company ever That that's me. And You think about it, right? But when you think about it, that makes perfect sense, right? Who would want to be able to get into work and start working on their email at seven o'clock in the morning? It's only someone who's highly engaged and really excited about doing what, you know, making an impact in the universe with regards to their job. So it was really interesting to be able to see that. And then this other thing that popped out of the data, which was really interesting, is people who work over 10 hours per day are less likely to quit than people that work eight hours per day. And that sounds counterintuitive, but when you really think about it, it makes perfect sense. People that love their job work on their job longer than people that kind of don't like their job, right? Going back to the boredom variable, it's like if you don't really like your job, if it's 4:59 PM and you're out the door, then of course you've got a higher chance of quitting. But if you're like, oh man, this stuff is so interesting and I want to continue to work on it, those are the people that you wanna be able to encourage. Now you don't want them to get burnt out, right? You wanna be able to say like, Hey guys, you need to leave the office or you need to stop working, um, at 5:00 PM or whatever your hours are. But we found it really interesting that this 10 hour per day variable was just so important, impacted The overall. Do you think this is where AI can fit into, like, I feel like AI, when if employees use it and as it advances, can remove some of the work, some of the things that make employees forward at work so that they can do more high impactful work? Absolutely. Uh, so like my long-term vision of where I think all of this is going is, do you remember Clippy the little Microsoft Windows? Yeah. Yes. Icon guy that would jump, I'm dating myself maybe, but like, no, windows 95. Okay, cool. So, and hopefully every other people, uh, people, uh, remember that too. So Clippy was, he kind of p****d you off, right? He was just like, I can help you with that. You know, I wanna help you do this or that. And they got rid of Clippy. 'cause Clippy was super annoying. What if you had the ultimate manager, clippy that was with you and was saying, Dan, you're filling, you spent the last 45 minutes filling out cells in Excel and I've kind of figured out what you're doing. Do you want me to just give you this formula to just auto complete it in 10 seconds? Like that is where I think we're going. When we think about AI for work, that is really valuable. And so you're going to be able to have this co-pilot that works with you throughout your workday to be able to tell you what you can do to be good at your job versus what you, you know, can't do to be good at your job. So you're only working on really difficult creative problems that mm-hmm, a human brain is good at versus selling, you know, filling out Excel spreadsheets is, uh, I don't know about you Dan, but I hate that. So that's the thing that I hate doing. And, uh, the faster that I can get rid of that and have an and hand it off to an AI, the better. So next one here is, uh, newer tech has a huge impact on impact, uh, huge impact on quick chances. So we rebuild our AI algorithm each month. So every single month we basically rebuild the algorithm and we collect the new data. 'cause we get thousands and thousands of people that quit their jobs, and then we add that into the algorithm. And last month we found that a variable moved from like the thousandth most important variable to the eighth most important variable. And it was only because we just collected enough data to be able to really see its impact. I, you know what, I'll do another one for the, the book, if anyone can name what website or application that's relatively new that we've been talking about in this presentation. What do you think, um, this website or application is that is so impactful? It's the eighth most important variable and it didn't exist six months ago. People are saying working, just name up chat, GBT Slack chat. Most, most are saying chat GBT, we have one bar Chat, GBT Google Would be upset if they only saw one bar. So if a worker starts using large language models like chat GPT in their job and then stops, it's one of the most important indicators that somebody is gonna quit their job. It's the eighth most important indicator that they're gonna quit their job. Uh, and again, this just kind of connects back to their passion and engagement of this particular position. They're saying, Hey, you know what, we can't actually, um, am I interested in the future of this organization? Am I interested in this business? Right? Maybe I'm not. Uh, this is something that is just so interesting for us to be able to see that you need to also analyze this data literally month by month, week by week, because there's so many new pieces of technology that are coming in that are impacting the successes or failures of people joining or quitting their jobs. So those are the six most important ones that, or the, at least the most counterintuitive ones that we came up with. Um, I would love to be able to get everyone else's perspective on where it's at. If you want to test this out, you can go to time doctor.com to be able to test it out. Uh, we actually have a beta program. So this quick prediction algorithm is connected to our beta program that we are, uh, that we have rolled out over the last year and a half. And we have all of these cool things, not only quick prediction, but we have onboarding prediction, we have management prediction, um, and a bunch of other really cool things like benchmarking across your industry, how are you doing in comparison to other agencies or other HR professionals in, in the space? So if you're interested in checking that out, just go to time doctor.com and uh, sign up and we'd be happy to be able to, um, take a look and help you out. But other than that, I'd love to turn it over to some questions if anyone else has any questions about where we're going. Yeah, definitely ask questions. The first question is, do you have any results as it goes across cultures, as in different countries other than North America? Do you see any trends or is it it's too early, It's too early to tell, but we do have Any hypothesis? Uh Oh, yes, I do have a lot of hypothesis connected to that. I would say, uh, there's a, there's a couple books that are written on this particular subject. Um, the biggest one is on the tip of my tongue. Um, the author spoke at running remote a few years ago. But essentially, I would probably say when we look at our data, the hardest workers are usually people in, uh, Southeast Asia. Uh, and we're seeing generally longer work weeks coming from that area in comparison to other markets. So that definitely has an impact on the data and we would probably build a separate algorithm specifically for that particular space. We did add in, uh, remote versus work from home, and we found some interesting data based off of that. Essentially the remote workers work week was 11% longer than the, um, than the in-office worker, but you have to also account for commuting inside of it. So one of our other, it is a longer work week, however, if you work in commuting into an office environment, that's not something that we can actually account for inside of time, doctor. So it's, that was an interesting, um, insight as well. And just the way that that data flows is completely different from the way that in-office workers work, essentially a lot more Zoom calls, obviously to be able to add into the remote worker versus the in-office worker, but then also, uh, a lot more distractions in the office environment. So we found that the amount of time that people were consistently in the zone working was much lower in an office environment than it was when people were working from home. So like someone could just sit on their word processing tool for two hours on average, whereas an in-office worker would be sitting on it for an hour and a half. Makes sense. Any thoughts on how AI use during work will be monitored and regulated in the coming years? I think that it's, I think that it's gonna be a opt-in not opt-out model. So my perspective is that we're gonna really see, um, definitely we're gonna see these metrics become more prevalent, particularly as it connects with remote work and hybrid work. And as I think we've talked about before, we've talked about this Dan, but there's a significant chunk of the US workforce that is working from home at least a few days per week. You need to be SOC two compliant, PCI compliant, HIPAA compliant to be able to make sure that if you're working with banking information as an example, that that terminal is secure and that someone just walking through your kitchen can't just like grab my bank account information and steal all the money outta my bank account. So we're going to have to have that just for security. I think on the productivity side, that's where there's a huge opportunity to be able to do things that we wouldn't have otherwise even thought possible five years ago. The kind of stuff that I just reviewed here as the tip of the iceberg, again, this technology that I just showed you is the worst it's going to be, right? Like, we're gonna get much more precise models, not only us, but everyone else in the space is gonna get a lot more precise. So I think it's really gonna come to kind of like a virtual HR AI professional that's gonna work with you to be your coach to say, how can I, how can I make sure that you don't get burnt out in the work that you're doing, that you find it interesting, engaging. Um, it could detect specifically when you're kind of done with working on one project as an example and suggest something else that you do inside of the organization. I see this coming down the pipe probably within the next five to 10 years for Sure. Utilization's gonna be huge. And you know, even at the highest levels, I mean, countries are having discussions about AI use, right? Because yes, at the highest level they're like, we don't want this to have a negative impact on mankind. I think there's like, you know, at least a dozen top countries that are in discussions right now. Plus you have, you know, major cities like New York that are trying to figure out, figure this out. And so it's a, it's a really, really big issue all the way at the top and that will, you know, we'll, we'll see more of that I think in 2024. Absolutely. Uh, we have another question. Any difference in quit rates and insights between remote versus hybrid? And I assume that they mean, uh, fully, like fully remote versus you obviously hybrid. Yes. One or more, yes. In the office, Uh, we have, we have run that split and we've analyzed, I believe it was remote, but included significant hybrid. So the way that we built that model is we basically said if you spend like one day in the office per week, we're gonna actually call you remote, we're gonna put you in the remote bucket, and then if you're maj spending the majority of your week in an office, we're putting you in the office bucket. And we found between remote and in, in-office, uh, remote had a 48% lower attrition rate than their in-office counterparts. So like pretty huge. Uh, but this data has already been clearly like this has been laid out by people like Nick Bloom, um, Roger from Harvard. They've all very clearly stated that when you work from home, your chances of quitting is significantly lower than when you work in an office. But this just reinforces that thesis. Yeah. And that also can be like negative and positive too. Yep. Go ahead Dan. Yeah, I, uh, I I've worked fully remote for 11 years or almost 12 years and my wife works fully remote and so, you know, I think it's something like 18% of positions in the US are fully remote. I believe that's the last stat I saw. And part of the reason why she would stay with her company is because to find another company right now, again, like everything's still being sorted out, would be very difficult and that the benefit of fully remote is worth more than her, than maybe a pay increase of, you know, a certain percent, right? Yeah. So I think that's another variable when it comes to, when it comes to that as well. Absolutely. Uh, there was a study from Buffer on the state of remote work report that they do every single year and take this with a grain of salt because the data set is obviously people that are excited about remote work, but they identified that 50% of that people, of the people surveyed would take a $20,000 pay cut to be able to continue on with their remote job, which is pretty crazy when you think about it. I don't think that's applicable to all workers everywhere, but you need to be able to keep that into consideration if you're working on your return to office, um, you know, documentation, you thinking about hybrid versus fully in office. To me the ability to be able to be open in your flexibility is probably the most important variable towards success. And there's, as I think we were talking about before we jumped onto this call, flex Index from Scoop actually has some really good data on that, showing that the ability to be voluntarily hybrid is actually the best of both worlds. So you get the workers coming back into the office to be able to build their culture days, but they're not necessarily forced into must show up on Monday or must show up on Friday. Those are the ones that had the highest employee net promoter score. Absolutely. Yeah. And they just released more data, I believe it was yesterday that showed that for the Fortune 500 companies, the ones that were allowed employees to be empowered in terms of when they came into the office, they were more profitable than the ones that didn't and mandate. Oh, really? Yes. Yeah. Uh, and another question is, do you have predictive data for engagement? Like what are the key predictors of engagement and Disengagement? So lack of boredom essentially, which was, as I said before, using their calendar app a lot, not using their top three applications. A drop in that usage of top three applications is connected to boredom. Uh, and then obviously the reverse. So in terms of engagement and non-engagement is the biggest ones that we've come up with. I think that you really, and these are just indicators to be able to have a conversation with employees to say like, what's going on? Right? Like the algorithm can really only show you something's going on here. Historically, these people have quit if they keep going on in this trajectory and here are the specific things that you need to address, but they could be 10,000 different things. The biggest thing that you can do is just jump on a call and and figure it out with them. Yeah. And there's a lot of different applications when we talk about predictive analytics, people analytics, the big one you show today is quit rates. I mean, that's a huge one, right? That's great for retention culture. All it checks a lot of the, the typical, you know, KPIs for hr Mm-Hmm. But what about other applications? Like for instance, you know, knowing when someone is well equipped and ready for a promotion, for instance. Yeah, so we do have those algorithms as well. They're less precise than quit prediction. But again, for us, it's just a matter of time before we collect this. Really, at the end of the day, the beauty of, um, of AI is everyone with, with chat GPT essentially AI exploded, right? Like if you look at the tech industry right now, every third company is an AI company, but the thing that they're missing is the dataset. And so moving forward for us, we have the largest second by second work database on the planet. So for us, it's relatively easy to be able to build out these algorithms because we have hundreds of thousands of users that, you know, give us data on how they're doing throughout their workday and then submitting that information to us and we can process that information. So, um, in terms of someone becoming a manager, there are some interesting indicators. We have some indicators connected to leadership as an example. One of the biggest ones is even inside of a Zoom call, how often do you talk inside of a Zoom call is correlated towards your ability to become a manager. There's actually, and again, this is like really early data, I don't even want to talk, I'll talk about it here, but, um, someone spending, someone being the second most vocal person in a zoom call is an indicator that they're ready for more leadership inside of the company, not the most, that's the person who talks too much, that you just tell them like, shut up, stop talking. Right? But the person that's the second most talkative person inside of those zoom calls is usually the person that's ready for management. That's just one little to all small tidbit of all of the things that you can add in. And again, we're talking about thousands of different variables. Um, happy to talk about that too. You guys can just email me and I can send you those data polls if you want. All right. Two more questions then we'll wrap up. I mean, I could sure talk about this for the next hour, but, uh, the first one is, how does that work for introverts who are not as inclined to speak in larger settings? I think that that's talking about promotion and leadership. Yeah. So I think that it's, uh, I, that's me by the way. I'm very, I'm very much on the introverted side of the spectrum and it is a central thesis of, uh, my book, which is if you go into an office environment right now, I don't even have to hear what people are saying. I will generally figure out whose ideas get adopted. Most often it's the six foot plus white guy that looks like Captain America. That's usually the guy whose ideas get adopted most often. Why? Because that person is really charismatic. They're very good at communicating that information. Maybe they don't have really good ideas, so maybe their ideas are like a three outta 10, but the packaging, their ability to communicate that information is a nine outta 10. So whose ideas get adopted? Well, it's a three outta 10 ideas packaged through a nine outta 10 person. Mm-Hmm. And so asynchronous management, which is the thesis of my book and why I think most remote companies succeed, is that when you remove those variables, better ideas have the ability to be communicated more effectively because they're not packaged through these Captain Americas as I like to call it, inside of organizations that are incredibly charismatic, but maybe don't have necessarily good ideas. And if you're a Captain America, and you're kind of p****d off at me right now, listen, I love Captain America's, uh, they're, our entire sales team is filled with them. They're, they're fantastic people. They're absolutely needed inside of an organization, but their ability to convince people of your ideas is not necessarily moving the business forward. So those introverted people are absolutely critical. And the way that you would do it is through asynchronous management, in my opinion. Yeah. And this, this comment is right on target. There's definitely a lot of biases when it comes to who gets promoted and whose ideas are taken seriously. And those have always existed, right? And, and you know, maybe those biases become part of the AI or maybe that kind of gets sorted out and there's more fairness in the future. And the other thing too that you can do as an individual worker is you can look at what is the ideal model for getting a promotion inside of the ai? And then you can coach yourself towards that. How often am I talking in meetings? What am I talking about? Uh, how often am I, you know, communicating with other team members inside of the organization? What messages am I sending through Slack? All of these variables can impact that model. So you can actually look at that model and say, that's where I want to go, and therefore here are the things that I need to change throughout my workday. I love that. And even from the employer perspective, you can say, Hey, we're not getting the right leaders. Let's look at what data we're feeding this, right? Right. Let's look at our criteria and then maybe we have to recalibrate. So I like that too. And I do like the individual aspect of, hey, like maybe it gets to a point where companies equip their teams and say, Hey, here's some data to kind of show this to you, you know, within reason. And then now they're better because they have that feedback. Absolutely. It's almost like an absolute, it's a new new performance review. Yep. But it's a performance review that's actually fair because it's the AI that's doing it. It's not, you know, your manager that doesn't really like you. And how often have we had that where it's like, I've got a bad performance review, but it's because my manager doesn't like me, it's not because I'm not doing good or bad work. Exactly. Wow, this was so much fun. Thank you so much for sharing this. You know, judging by the comments people are, people are really interested in this space. We all know, I mean, it's not like data's gonna be used less in the future. It's pretty obvious. And, and this is definitely one way to be able to extrapolate and put models together that actually improve talent outcomes, which is what everyone who's calling in is interested in doing, right? That's, that's how people are getting evaluated as part of hr. So I wanna thank you so much for your vast knowledge for giving us such a sneak preview of everything that you're doing at time, doctor, with this incredible tool, uh, for everyone. You know, you can go to, uh, you know, on the slide, you know, get dot time, time doctor.com/beta-program to be part of this. Uh, you can also go at Url, but yeah, just time doctor.com is probably easier. And then running remote.com, uh, we'll get you to register for the next big event they have coming up in April in Lisbon, Portugal. So any, any final remarks? Uh, Leah, No, this is so much fun. I, I love to show people this stuff and as I said before, this is unreleased data, so it's great that everyone's been able to give us really great feedback and I'm excited about building this out in a deeper way. So if anyone's interested, again, hit us up time doctor.com. I'd love to be able to chat. My email is liam@timedoctor.com, so if you have any questions, even if you want me to just send you these slides or anything like that, just email me and I can do it. Excellent. Thank you so much and thanks to everyone for joining today.

‍