How Can AI Know You're Quitting Your Job Before You Do? FT. Tyler Hochman, FORE Enterprise

Can AI predict when your employees are about to quit—before they even realize it themselves? Tyler Hochman, founder and CEO of FORE Enterprise, reveals how his company uses artificial intelligence to pinpoint hidden workforce attrition risks and proactively tackle them. Learn why seemingly trivial data, like where an employee eats lunch, can be a shocking indicator of turnover. 

https://foreenterprise.com/

Transcript

0:18

hi everybody welcome to another episode of the AI Rebels Podcast we're very excited to have Tyler Hoffman on with us today founder and CEO of 4 enterprise for among many things as we've recently we were talking before the show you're branching out into so many different things it originally started right as more focused on HR attrition really digging in and using AI to tackle this problem is that right yeah okay perfect maybe Tyler give us a quick give us like your elevator pitch of 4 as it stands now what is what's the the big idea here yeah I can I can give you the elevator pitch about when we started and then about a year ago we made a a pivot um and I'll give you the new elevator pitch the first one is that we could predict when a businesses employees were gonna leave well before they themselves even knew they were going to leave uh and as a result of those predictions we could then uh create targeted intervention strategies in order to prevent them from leaving or to to optimise around when when and who is going to leave uh and and then a year ago we pivoted and we still do that but we also now are a solutions architecture firm that builds the end to end data pipeline necessary for tools like that workforce tool but uh also for other AI enabled tools interesting what LED to that pivot good yeah so a lot of our uh clients you know would would would ask us would be really drawn by that initial workforce technology and then we go in and we say okay you know hand us get give us the data cause obviously there's a lot of data you need for that and in that handover we'd find out that most of our clients being large Fortune 500 companies hospital networks you know did not have the necessary the data pipeline necessary to actually accomplish those analytics yeah then we're like okay well everyone's so then they say why don't you just go and build that for us we're like okay we'll do that first and by the time you finish building it you know we're a year into the contract and we're like why don't we just build this for people oh my gosh that's you're like great we'll uh sign you on for years to come ha ha that's a lot of work though okay well maybe let's back up then and go like what LED to that initial interest was it your interest in that attrition HR what LED to that yes one of the businesses I started in college was a quantitative analytics firm that service financial institutions and what we found was that when companies we looked at a lot of like m N a deals and we found that when you're trying to value a company one of the things that's being discounted or not correctly accounted for in the market is the true cost of turnover and the cost of workforce sure company so as a result of that when I set out to kind of build a company or you know build yeah build this company we I wanted to address that and I wanted to help businesses understand not only the true cost of turnover but who is going to turn over and then how to prevent that turn interesting so what has been kind of like the the most unexpected variable that you guys have identified as an early warning sign have have there been any any that stand out or has it all been sort of more confirming what your pre existing suspicions were you know there there's a lot of like more technical variables that have to do with the creation of different metrics um and like this the deviation of different metrics you know there's obviously the funny ones which are like you know if you uh some some companies actually know where they're and and again and I should add caveat this was saying that the only data we collect from a business is data that the employees already know we're being collected so we don't actually go and so as an example if a company tells you they're not reading your emails we don't go and read your emails like it's only data that's already being collected but some companies collect data like around you know where are you taking your lunches as an example if if you're on a campus style environment and you actually have to swipe your card and order lunch from these different places so there was some correlation in in a in a company with a campus style environment about where they took their lunch and whether or not they're gonna stay or leave them oh my gosh that's super interesting that's wild that's crazy so what is it like what's the can you walk us through from like high level start to finish you come into a company they're like hey we have this attrition issue we wanna figure out what's causing this and then what do you do like what's that process yeah um so a company will come to us uh and say we recognise we think turnover is a problem most companies can't even define like the true cost of turnover which is anywhere between 60 to 400% of an employee's annual salary you know depending on the skill level and the amount of onboarding involved it would be it would be like if you had to replace Spencer right for this podcast and mm hmm really hard to find Spencer again yeah impossible impossible impossible especially this hell yeah I mean that would take years that would take years to replace that exactly so really expensive problem um and uh so then we the first thing we do is we actually help them define like what is their cost of turnover okay that cost we normally start depending on how big the organization is with a pilot group um you know anywhere between 100 and 200 people or a department or you know a building or so on and then we go and we you know collect the data and we create a dashboard and on that dashboard you can see it by team by department by you know whatever groupings you want or by individual and then it assigns them a ranking system of you know how likely they are gonna turn over from low medium medium medium high high and very high um and then corresponding once once it once it passes a certain threshold like high to very high we actually tell them the reason we think they're going to turn over and what to do to prevent it super interesting so in in mapping out what to do to prevent it what kind of work has gone into figuring out those solutions for customers has it all been just historical data from previous customers like you know tracking tracking the the individuals that you said we're going to you know leave and then we're able to be retained or what what does the process look like there there's kind of three ways we do it uh the first way is if the company has tried intervention methods in the past and actually collected the data on those intervention methods but never analyzed it that's probably the best place to start is you know actually doing that analysis and saying okay you know as an example we work with a hospital network who tried giving a meditation app to nurses um and we're convinced you know oh you know nurses are saying they're burning you know our nurses are saying they're working too hard why don't we give them this app and then they can meditate now they're not now they're gonna be happy again I did not work yeah it was not shocker yeah shocker there I was not and so we you know so that that's kind of the first place we'll start uh the second place we'll start is we have a really cool group of advisors um from uh Stanford and Wharton and Kellogg and all these different places who write a lot of research uh and on this topic um and so they're they're a really good starting place for we consult them and ask them you know where where we think this this problem of attrition is is occurring and things to solve it um then the third place is just training our algorithm on kind of historical uh attrition solutions typical attrition solutions like as an example the most well studied one is the recognition problem uh and people found that just recognise an employee create a three times reduction in their desire to turn over um and so and that just means like saying good job you know it's literally as easy as that three times reduction that's crazy brings to mind the the old office joke from Michael Scott where he where he's talking about like you know how somebody goes into work doesn't get an award and you know kills himself and it is it you know that that it's always presented as as a joke but it is funny when you didn't dig into it yeah how much how much that stuff matters I was training a customer support team for for a while for my job and yeah I found similar similar things where the the people who would stay the longest were the ones who you know we interacted with the most and that's really what humans are looking for is like those interaction and social validation that's fascinating I'm curious so oh go ahead Jake I was just gonna ask um along these lines cause obviously you each company's unique you go in you get their data you train on the you you know you eat their data and you decide what does this look like now what does it mean for you at some point do you think you'll have enough where you wouldn't even need to do that like you're like oh you're this industry you're this size of company you gotta fix these problems you know what I mean like could this turn into more of a almost like you have a consulting arm as well it's less the data more let's just help you implement these best practices that we've discovered yeah it's a great it's a great point we can already do that and we do do that for some businesses like we do diligence for certain firms when they're looking in acquisition and so we'll never actually get you know the true company's data but with enough information industry number of employees location all that kind of stuff we can do that you know just it relates to our confidence interval uh the 1 the confidence interval of the prediction and then how it changes as you go multiple months out right so it's actually a lot easier for us to stay and employees going to leave next month and it is to say an employee is going to leave you know then try to predict the more months you go out so the employees gonna leave you know got it four months or five months or six months um and so the internal data gives us that breath and to be able to say six months out we're already identifying employees that are leaving versus you know just you know a closer window and it's all related to like a confidence interval so we could still even without the internal data predict on a six month or one year to window it's just yeah confidence is gonna be less and I'd be interested to hear um are there industries that you have found are are easier to predict than others or are they all kind of about the same once you have either the right data points oh no there's definitely I mean it it normally correlates to like how much data is collected on that industry you know well regulated industries or well right I mean yeah regulated industries have a ton of um healthcare finance are the two that come to mind but then you even have like certain you know government related security jobs government related space jobs and space industries it's you know very well regulated industries just mandate a ton of data being collected yeah yeah makes sense interesting have you had any issues with companies I imagine like I remember in high school reading the scarlet letter right where you have like the red a on your chest like is there any impact where you come in and you say like that one that one that one they're all leaving in a month and the company have you had any issues then like signing off these employees almost and like treating them poorly rather than help trying to help them stay has that ever been an issue yet you know I think you're raising a valid concern which is once you know whether someone's gonna stay or leave that additional information I'm certain does impact you know the way you're going to approach it because you just didn't know before um yeah to the point that it makes them treat them worse so to speak I'm I'm pretty confident it doesn't just because the status quo is they're leaving so yeah if anything you're not I see you potentially run into the problem of not treating them better but you certainly are not gonna treat them worse because if you want them to leave and they're already leaving you're just going to continue doing exactly what it is right yeah I imagine like like this I this employee is gonna leave in three months and I just have this scene of like you know you you're pretty confident that they're gonna leave but they don't even know it yet and you're like you dirty like I know you're gonna leave just seething in every single one on one I'm so mad at you they're like when your wife has a relationship dream about you cheating on her or something like that yes like just mad at you for a week yeah yeah exactly that's awesome well I guess that brings up an interesting point like cause this is a concern a lot of people have with a I is at what point are we like are we morally like what are the morals here with being able to see the future you're like like I am 90% confident you will be leaving I think that scares a lot of people that AI can do that how do you address those concerns when I don't know if anyone's brought that up but I'm curious what you would say to that no no it it has been brought up um you know we have metrics in place there's there's two kinds of metrics safeguard metrics we have in place the first one is like we always wanna make sure that our a I isn't contributing to systemic biases you know with a company um and so we and then this is kind of also to the second one but I think ultimately and to answer your question and then come back to this which is it's information right like information can be used positively or negatively but I am always a fan of more information is better like you should never hide away from the truth you know yeah you can choose to confront the truth either way but like facts are facts information is information yeah and so that that's kind of how we approach it and so we create the safeguards but yeah you can use information you know each each way you can take a good or bad yeah yeah crazy world it's just yeah I mean it's so what what kind of go ahead sense you brought up bias and so I thought that was an interesting segue to asking you what kind of are there any specific safeguards that you that you use to reduce the bias you mentioned that you know you do use some certain safeguards to reduce systemic bias in companies are there any of those that kind of stand out in being interesting to talk about I guess it's the question I'm asking yeah I think uh I think there are for the usual biases you know that that come from like race ethnicity socio economic status you know location those types of things um I'm trying to think I mean there's the age bias I think is always fascinating in tech um and one that I actually think is not talked about enough going back to the office there's a great office episode where Michael is gonna sue Dunder Mifflin for ageism right now I think it actually does apply to deck so you know come full circle over here you know that's kind of true yeah yeah that's a very good point and the safeguard you know just generally involves actually having real time metrics like again our metrics you know aren't necessarily the the biggest thing we're offering is predictive versus like retrospective and so it's not like you can't see who is going who has turned over and then look at the relative associations with that turnover it's just that now we're predicting for the future and making sure that you're staying ahead of those biases makes a lot of sense I love that you've mentioned data a lot obviously that's the the life blood of all AI how do you determine when you come into a company how do you determine good data like how do you know if this is gonna be worthwhile it's actually it's actually a really nice segue into the fact that like that's kind of the reason for pivoted away from just workforce and focusing almost more on on the solution architecture and the pipeline itself because businesses right now are being tasked with like you know a I Fi my business you know and everyone's trying to create these really cool a I tools you know in order to do that and both from a B to B perspective a B to C perspective and I think what's being undervalued is is the data right it's it's the way that data has to be ingested it has to be structured and streamlined it has to be stored MHM and then it has to be ready to then power these a I enable tools yeah yeah and it's you know and not only is it being underestimated from a you know labor standpoint it's been underestimated from a cost standpoint like it's still very expensive to accumulate large amounts of data to store that data and so yeah I mean you know open AI only works because it's literally burning billions of dollars I mean you know they've got deep sea got these other places competing with it but it's literally burning money like it is not a cost effective solution I mean close as it stands right now and so I think people that's where that's where there's I've I've noticed a disconnect and actually we turn away a lot of business because we're like this does not make sense for your business you don't have enough cash flow it's a really an expensive endeavor to do it or an expensive firm and you know it it just doesn't make sense you know it's it's too expensive for your business yeah yeah I think people really do I mean data if you think of the way I think of AI sometimes is like a brain right where if you took a human and you took all their memories away like would they even be could they do anything not really or and then if you even got like less extreme right if you just like took out some key memories throughout their life like they just would become less and less authentic and less humid and I think it's the same thing with AI where like if you if you don't have good data that's the AI's memories that's what that's its brain that's all it knows and so if you're trying to companies want all these solutions we're we're seeing this so much we talk to all kinds of companies right and everyone's having this issue with data it's like look really cool idea for a tool you can't you can't do it like you don't have the data here to build this which is discouraging I'm sure what do you tell come like do you just say no or do you kind of say hey this isn't right for you now but start doing these things with your data and maybe in the future like what would you tell a company that was struggling with that I would say that you know it's a function of cost data is acquirable uh so it's just you either are you you have it's a zero sum game with dollars and so you have a finite amount of resources to allocate in different ways assuming you have a revenue base and a cash basis that I could even support it you know in the first place the first place I was like do you even have enough cash basis to support this type of endeavor if you do is this how you wanna allocate your dollars and if it is yes we can go and acquire the data most of the time necessary to build what you want um that being said like it's you know even something as simple as a chatbot right that is specialized to a specific business it still requires you to go and actually vectorize a database you know put all these embeddings on top of it and then you go and you have to you know store it and host it and then you go and actually use it to right right the chat bot and you know most people don't know that they're just like no I want this thing that talks to me I want it yeah service questions like don't know if you understand totally yeah hundred percent so I'd be I'd be curious to hear could you just kind of walk us through the flow of of how your tool works a little bit yeah yeah so we essentially have found created proprietary ways to ingest data are you talking about our pipeline tool or our workforce tool let's hear both yeah let's start with the pipeline yeah since that's that's what we were just talking about yeah yeah yes on the pipeline we create a proprietary way to ingest data at like what we think is a pretty cost effective you know caught within a cost effective model and that uses all different types of techniques like sampling techniques you know clustering techniques and so you can actually ingest less data than than you were initially thinking you were gonna need to ingest then from the structuring and streamlining perspective we use a lot of a I there actually think if you ask me like Tyler what is what is the thing that you think a I does best right now my answer would be in structuring and streamlining data it's taking really it's doing mapping right like I think you know this agnostic stuff there's a bunch of stuff coming out that's gonna be very cool and what it actually knows to do best right now is like is like mapping you know you give it thing and then you tell it you know to make it this this thing look like this thing and it's really good at doing that it actually does that really well most of the time so that's where we use a ton of you know different AI and AI models then we have a proprietary way of storing it and hosting it so it's cost less and then ultimately we feed it into you know any type of tool and then it's an example of one of those tools would be a workforce tool that we have um and that tool intakes the data it goes through our different predictive algorithms that take into account all these different input variables both internally from the company but also externally from everything from census information you know to demographic information all that kind of stuff um and then we output that dashboard that I was talking be curious to hear do you think that there is cause obviously right now a hot topic in AI in general is synthetic data do you think synthetic data could ever have application in something like this that is so reliant on you know real world historical data and and as a follow up to that if so do you feel like you have hit the point where you might start doing some of those experiments yeah man I from I would say you know synthetic data offers very interesting opportunities at a cost perspective one of the things like we champion is is trying to get that cost down for a lot of our clients as as much as we can so I think it's very interesting from a cost perspective in terms of have we hit that point yet I I you know it's it's definitely something we've explored especially as you try to build like front end solutions and you wanna test them you don't have the time and the resources to go and actually allocate to getting the data um so yeah it's it's something that we've explored before interesting with you mentioned this is kind of going back but you mention you have all these experts and scientists from all these leading organizations involved here has there been any really heated debates that people have had about we need to we need to do it this way we need to train the I AI differently it needs to focus here anything like that Jacob you just you want the drama don't you you want you wanna spill the tea bring it out you want you want some drama here ha bring it on yeah well you know what's funny too is like a uh you know a heated debate between two professors and like one of our lead engineers is the calmest debate you've ever seen yeah yeah anchor chief might be the most affirmation you get right affirmation you get yeah I mean you know there's there's I think there's a lot of debate going into uh causal analysis right now we've consulted with professor at Stanford who you know just wanna a Nobel Prize in economics for it and it has to do with understanding like like you know functionally the root cause of the reason and in in our use case would be the root cause of why someone is going to turn over and then you know the actual intervention strategy to preventing them from turning over um because there's so many correlated variables uh that it's it's just really hard to say and so that's one of the things I think we do well is doing that analysis to really understand this is the attributable variable and not you know all these clusters and and noisy variables was there a key key like something Learned from that professor that has really impacted your AI algorithm like what what did he tell you that helped you zero in on the crucial causal factors there yeah I mean it wasn't you know it's it's not anything flashy it's more like technical ways to scream out noise yeah yeah no I don't think it's the drama you were looking for but yeah no that's great that was you know a lot of the insight we got yeah yeah well yeah I'm sure that's very complicated we had on the show this other organization who use AI to basically take world events right the eggs eggflation right egg prices going crazy they take that and they use AI to dig in go explore the entire internet and figure out what actually is causing it right if you go to CNN it's gonna say something really complicated ontologies and knowledge maps it's super cool huge but it's like I mean you just have no idea what it takes to figure out the cause of something let alone when humans you know when humans get involved like this like I'm amazed that you're able what kind of a I guess I don't think we've even asked this what kind of a result do your customers see in attrition levels so they see what we describe as like optimized attrition levels which you know natural attrition is good uh you actually want people coming and you know coming stay and leaving a business at the same time you don't want runaway attrition and you actually don't want too low of attrition as well you know because then you don't have that natural attrition anymore yeah and so they see an optimizer so they essentially get to they get to pick how much attrition they want assuming they're willing to do the intervention I'm curious have you seen like a kind of a general level of attrition that that most people sort of settle on or is it sort is it pretty unique to each business it's pretty unique to each business like if you're you know a business that's in a period of rapid growth as an example you you pro you probably will have a lot of attrition just because as a result of growing so quickly you naturally are not not a stable organization that being said you don't really want a lot of attrition because you wanna maintain your workforce to grow faster and then conversely on the other side if you're a well seasoned organization you may actually be you know with plans and processes in place the cost to onboard someone might actually be a lot less expensive so you're okay with some mathematician but you actually don't have that many or you know and so ultimately it you know it just depend it really depends on where you are you know what industry you're in all those different factors wow okay are there are there any other pivots in the future or you just digging into this one for now cause that's big that's a big shift we're digging we're digging I mean ultimately this one became like a priority for us to even do the workforce analytics and so we were more than happy to to kind of be here and we we get to what's what I love about it is like I have a natural urge to I've started a number of businesses now and and to kind of continue to start businesses but what's cool about this one is data pipelines are look the same no matter what business you're in so we actually get to touch a lot of different businesses everything from we work with a fashion startup too we work with financial institutions we work with PR companies yeah we're getting to touch a lot of different things yeah that's fascinating I'm curious do you ever work with any anybody in like like the fast food or the restaurant industry that so we have not touched the fast food or restaurant industry in a really in depth way we've done due diligence projects on those one of the interesting results we found that was in conjunction with another project we did for a large retailer is the true cost of part time turnover which we found as it related to this this specific retail organization was more costly than on a percentage basis than their full time turnover which was a really interesting insight for them yeah that's that's part of why I asked cause cause I was just thinking about it like that you know that's a that's an industry with a ton of natural turnover already just cause the jobs kind of suck to work so I was curious if there was you know anything else hidden in the numbers that's that's really fascinating here I I wouldn't have guessed that actually I would have guessed that their their processes are generally optimized enough now to turn through employees the way they need to right right what and it's also you know again it's on a percentage basis so like the output if you the and the reason for it is like if you can get decent output from a part time person on a you know cost utilization benefit like the cost is really low and so as opposed to a full time person and so when that person turns over although the total output might be less than a full time person for the good ones at least you know because you're not paying them that much it really hurts and so it was it was yeah the part time turn it was very expensive for this organization that's super that's that's fascinating yeah it's very interesting Tyler if it's not a to proprietary from all the data you've analyzed if if you could give one to three suggestions to any company if they wanna decrease attrition what advice would you give well yeah I'll give you I'll give you even the the way to the the most qualitative factor we've seen first on how to identify it and then how to address it um in terms of how to identify it the most and again there's no one answer I wanna preface it with that you know the reason right right that a I exists is cause it has to synthesize a ton of data but the most quality factor would be your schedule variation and so if you looked at the time at which you do any task right come or leave work open up you do your first do your second task whatever they may be and there's a ton of variation in in that on a day to day or week to week or month to month basis that's a very we've we've noticed it's a high indicator of turnover what to do to retain that person is to I mean the strategy that the blanket strategy I would give for anyone is what we talked about earlier which is the recognition strategy you know go and sit down and do a one on one and say you're doing a really good job you're valued at this company it's shocking how much that will decrease a person from leaving the second strategy I would give is a scheduling strategy and that most people most people you know businesses have a lot of things that they consider and so one of the ones that is sometimes underconsidered is the schedule of the business optimized on a individual basis right and and obviously certain businesses it's easier to make it more optimized on Andrew basis as opposed to other businesses but I do believe that every business should still try to do it in in part now whether that even means you know am I offering a lunch break at the right time you know is my lunch you know does someone like to eat early do they like to eat late you know actually fine tuning their schedule to allow for that really goes a long way with me so to your point I'm sorry I was that Spencer just a quick anecdote to your point I I have the problem of perpetually being late no matter where I go so I was working this job where I was always consistently just 10 minutes late to my shift and my my boss was sick of it so one day she sat down with me she's like alright listen we're just gonna push your the official start of your shift back 10 minutes solve the problem but then were you 10 minutes late for the 10 minute late start time no actually I I I just kept telling myself you know that the original shift start time and and that worked I know I was gonna say you should have just told her you shouldn't have told me that you should have just done that for your head yeah exactly you guys should have done that on your side and let me out of it yeah okay well I mean that's very interesting shocking like how simple some of the solutions are with like just talk to talk to the people was I understanding right maybe it's just me being slow here were you saying flex so like if a company has lots of flexibility so you can do the work whenever you want right let's let's just go extreme like you can do your own schedule would that do you think that would decrease attrition or increase attrition it would decrease attrition substantially okay okay so that flexibility for someone if someone can do the work when they wanna do the work decreases attrition typically yeah exact and and again it doesn't even have to be as drastic as you know some person works at morning some person works at night just as simple as you know you're like to Spencer's story right you start 10 minutes later or your lunch break

instead of being at 12 it's gonna be at 12: 32:09

30 so you like to eat a little bit later in the day you know it's like my new and that's why I actually think for when we what was so appealing to a lot of businesses in the workforce space is it's the intervention stuff is pretty hard like if you especially if you try to get like very custom tailored interventions yeah yeah it's also expensive what was cool about 4 which most businesses can't do is actually identifying someone before they're gonna leave like we believe it's not that difficult to keep someone you may still lose them you may try you may do all our strategies you may still lose them what's difficult for most businesses is they have it's called silent turnover right they have no idea someone was even thinking about leaving and then the person just leaves and then they're caught off guard and so that's what we were trying to do is give the knowledge and the info so that doesn't and then the recommendation strategies themselves were highly effective but you know you could get pretty far with just a general recognition or schedule tailoring strategy yeah yeah yeah no no it makes a lot of sense um I like it a lot because it's uh having been in both like you know position of an employee and a manage you know and in management um from my experience in management from the management side a lot of the problem was not that like you know you necessarily didn't know what you had to do for a particular employee but more that like you yourself didn't have the bandwidth to do anything so it's it's really it's really cool to see solutions arriving in the marketplace you know that that and it push those responsibilities to higher up to you know to people who whose bandwidth is dedicated to solving those problems and you can you can you know focus on on other aspects of your job I think that's really powerful that is if we go broad I'm curious how you mentioned how you started we always do a little snooping when we have someone on we go look you know and you've started lots of companies you've been involved with so many different things obviously you have a knack for identifying opportunities four is a great example of that you started with the HR and then you're like oh here's an opportunity we're gonna dig into this data is there something on your mind right now or something you've thought about where you're like man AI applied to this would crush it or would do whatever like is it have you thought of any ideas like that recently yeah we you know I can't I can't name the the in the sport or the team but I can just say that in sports there's a lot of cool opportunities for a I um both from a you know scouting perspective from identifying talent before it reaches maturity and then from a developing talent perspective and then from a optimizing for victory and for wins perspective yeah we're we're doing some cool work in sports um what I can talk about that's really cool is we work with a fashion company that is creating an a I stylist um and so you know if you know kind of pick out the optimized beanie for Spencer um would would be this company's for today let me know yeah let me know when they launched probably just send you the same green quick silver beanie you already optimized it man yeah exactly um yeah so we're seeing some cool stuff there we're seeing a really cool stuff in the MNA markets as well where we work with financial institutions in identifying companies that are you know right for MNA activity you know these places where large amounts of data exist and this this is of it can't be done by an individual person yeah cause hearing you talk about this oh I was just gonna say I have wondered to the the note of sports I've wondered how long it'll be until sports start making regulations around use of language models during the game cause I mean it's not it's not gonna be that long until you know somebody sitting in the box at a football game can be pointing a camera at the field and in real time have you know have an LLM telling oh that's a blitz that's a blitz package you know or you know cover to right like it's not gonna be that long and I wonder what kind of what kind of you know sport and cultural regulations will will develop absolutely yeah and also how do you even regulate that you know like how do you track yeah it's gonna be fascinating it's gonna be like I don't know if you guys are into the or you've talked with people into the whole like education AI education space but but I find it super cool that the way you know AI and L L m's are assisting uh you know studying and and learning and and then function cheating and then how people right develop like detectors of that cheating and then you know and then someone's trying to develop a humanizer for the thing and so it's like a war right it's like yeah it truly is look like yeah it truly is and it's like micro micro wars going on in like each industry yeah and it's also fascinating to see you mention like education we had a professor on and it's interesting how people also just like get around it he's we asked him like what do you do like with a I cheating he's like I just tell him to use a I like use it you know like more power to you use the tools available to you it's like there are a few things where I'm like you're not gonna have it in the test so you need to know X y Z but he just kind of beat the system by not playing he's like just use it but yeah it's gonna be interesting more and more like the regulation all the spaces I was curious you mentioned all these different applications sports um all these different things because hearing you talk about 4 and the way you've applied it to HR and a trition seems like you've just kind of built a model that you could replicate for data ingestion causal figuring out the cause and here's a prescription like do you do you anticipate duplicating this now with like okay here's our HR package here's our sports package for talent acquisition you know what I mean like is this something you're gonna keep doing yeah exactly I think that you know and then if you describe it more holistically like all the way from the pipeline to the tool itself you know that's exactly what how the business has evolved and what we're gonna go do and and we can do it you know we can do it across a bunch of different industries it's it's pretty industry agnostic so I'd be curious to hear obviously the the creator economy as much as I kind of hate that term it's you know it's it's taking hold it continues to grow Mr Beast is a corporation right like he's got a whole get up behind him um have you ever looked into doing work in that industry at all I'd be curious to just to you know here if you have it if you have it there's anything interesting anything unique about about it yeah we work with a large PR company so it's kind of adjacent to that industry because they track a lot of you know what the what creators are doing and how it affects brands and uh you know we we've created some pretty interesting recommendation models for you know this company to use with brands where they can recommend marketing strategies and creator strategies for brands to hit their certain goals um but yeah I mean I think you know that that's pretty pretty much as close as we've touched it but I think there's a lot of room I mean uh Spencer or sorry Jacob was telling me earlier how you develop the uh an AI agent for it seemed like for a meme coin yeah ha ha ha ha ha ha that was fun to do ha ha ha ha that was a good time crazy oh man let me think okay I actually wanted to ask this I can't believe I forgot this we hear obviously it's everywhere it's in the news it's everywhere AI is gonna take our jobs we're not gonna have jobs in the future everyone's terrified your company's positioned in a unique place where you're like literally helping people keep their jobs with a I what are your thoughts on the future of a I humans working is it actually gonna take jobs is it gonna create more than it takes what do you think yeah I mean you know I think I I I equate AI with something synonymous to in the most like fully flushed out example maybe the telephone um you know communication that that was previously took place by of in the old days by like pigeon and then by actual people ha ha um you know and then ultimately by a technology now we've kind of seen the full effects right of a telephone you know I am sure people were displaced by the telephone that being said I I didn't actually view it as like this majorly disruptive tool to the workforce and I kind of view a I the same way are people going to be displaced by a I absolutely but I don't see it as you know at least right now as a huge as it is offering efficiency gains to businesses and so businesses can now reduce their workforce as a result of making each employee able to work faster and smarter and harder that will that is all that is actively happening is it like completely you know like the way that Uber kind of got rid of taxis like no I don't think it's to that effect yeah it'll be interesting to see the the long term effects in anecdotaly I've seen I've seen a lot of people use it to start their their own side hustle whether that's just you know they've used it to bounce you know bounce business ideas around until they find one they that they wanna try or it's you know something surrounding AI specifically um and I'll I'll be curious to see if that's more like just an early mover type of type of um dynamic or if that's something that we'll see continue to replicate itself um as as it becomes more and more widespread um yeah it's fascinating fascinating yeah yeah I mean and to that effect you know like I heard this recently and I don't know how accurate it is but I thought it was interesting that someone told me that like dollars have never gone to zero as fast as they do now as a result of a I amongst other things also culturally people just trying to build businesses so quickly like everyone's trying to build a side hustle or a business and as a result of that the evaluation of a business like businesses have money and then don't have money at a faster rate than has ever existed crazy yeah or vice versa it didn't have money and now have money and also you get this funding yeah money is flying all around both ways negatively and positively yeah exactly it's it's crazy man well as we wrap things up I'm curious you've started companies 4 has been extremely successful we have lots of listeners that are trying to do what you've done they're trying to find an idea build it integrate like use AI in this process an AI tool whatever it is what advice would you give to these people that are trying to to build something like this yeah there's kind of two two pieces of advice I would give the first one is uh I think to start a successful company in general you have to be what I describe as like irrationally confident and so you will have a lot of reasons that it won't work most and by the way like the irrational part is that if you tried to justify it on like a board the con column would be much larger than the pro column but you still have to persevere right you still have to be like I'm I'm gonna do this just cause you are confident in your ability to execute and then the second one is to identify a problem that's extremely annoying hmm like in your own personally in your own life in your business life in other people's business life that you just find this like problem persistently annoying and it could be as small as you know change adding a leg to a table and as large as you know adding a chatbot to the world's largest service center because every time you call in you have to wait 30 seconds instead of having to wait five seconds right and so something that you just find very annoying and this personal to you I think is a great place to start a business love that that's great advice awesome well thank you for coming on if people want to follow you and for what are some socials what are the best ways for them to to keep keep yeah we're we're at for enterprise.com um and then we have a LinkedIn on for enterprise.com we keep our news up and all the things that we're doing we publish case studies and and things like that and then on our LinkedIn we we keep that active as well so would love to you know connect with anyone who's interested perfect we'll drop links thanks for coming on Tyler thank you thank you guys