Latest / Exit Algorithms / Why 95% of Companies Fail at AI (And How to Be the 5%) with Derek Crager (#82)
Transcript
- Pete Vera, Exit Algorithms: Welcome to Exit Algorithms, the podcast where we decode what it really takes to unlock growth, streamline operations, and prepare your business for a high-value exit. I'm your host, Pete Vera, and today I'm joined by Derek Krager. He built the highest-rated employee training program in Amazon's history and is now the founder of Practical AI and creator of Pocket Mentor, the world's first voice. based AI mentor that captures 30 years of your best people's expertise and delivers it through a simple phone call, eliminating downtime and boosting productivity. Derek, excited to have you here. Welcome to the podcast. Derek | HumanFirstAI.net: Well hey Pete, I'm excited to be here and thank you so much for the invitation. So I'd I'd love to pick your brain as you pick mine and let's I'm I'm exp I'm expecting a good conversation between us. Pete Vera, Exit Algorithms: Yes, sir. Same here. Yeah, I I love your your background and and what you're doing these days. Can you share a bit about your your journey and you know, business background? What made you take the leap from, you know, your s safe corporate job to starting your own business? Derek | HumanFirstAI.net: Well, I working with Amazon is never a safe corporate job, as you can probably hear about all the layoffs that's going on now. I I wasn't affected by those layoffs, I left on my own accord, but you know who knows what would have happened if I went in another day. So my background, I actually came out of school and and I went into industrial trades and I help build factories for companies like Ford. Chevy, GM, Chrysler, Subaru Suzu, that sort of thing. So my background is it falls under reliability, maintenance, engineering, that sort of thing. And I spent about 15, 20 years in the automotive manufacturing point, took about 10 years off, built my own local business that became the largest of its kind in the state of Indiana that was real estate related. And then I went back into corporate model, back to automotive, and then I finished out my corporate term with Amazon. I was diagnosed at age 50, called late stage diagnosis, simply because you know, 50 is not an early bird. And I was diagnosed as autistic, ADHD, and dyslexic. And what that gave me. really at the early point of my Amazon career was was validation or realization on why life had been so hard for me up until that point. So mm looking back, I had always been successful when I had the right information at the right time in the right context. That's what I leveraged when I was at Amazon, and I learned that my my diagnosis helped highlight why I'm good at certain things and why I'm not good at certain things. And ultimately my takeaway of my time at Amazon was reinforcement of that, that when the right people have the right information at the right time, they can accomplish amazing things. And looking at ways to augment and support other businesses. I had people tapping me and say, hey, you're at Amazon, you want to come work for us, which it's great. You know, you you get that headline from a you know a big company like that. But what I found out that not everybody has Amazon's money, if you can believe that. So AI, I l learned, can actually replicate the infrastructure that I had at Amazon for pennies on the dollar. Pete Vera, Exit Algorithms: Yeah. Derek | HumanFirstAI.net: And saving thousands of t hours of time that went in manual labor to do the same thing. And how I use AI, my company is called practical AI because it's really that practical usage of AI. not trying to throw AI at every problem out there. And so a lot of companies try to create problems just so they can throw AI at it. but what I came out to find is that Pocket Mentor really does offer the right information at the right time for employees, making those employees amplified so they can do their job much better. We're not throwing the baby out with the bathwater. It's humans that made these companies, gave them the value they have today. So let's not get rid of humans. Let's just augment the humans, amplify these humans with AI, and we'll have the best of both worlds. Pete Vera, Exit Algorithms: Mm. Yeah, love it. No, I love that story or journey. definitely unique. and thank you for sharing, you know, about your autism or dyslexia, ADHD. I think those are things a lot of a lot of people can relate to, you know. I I'm curious how how has that shaped your the way you build products and and lead teams? Derek | HumanFirstAI.net: what we found out is back in the old days it was one size fits all and you had to live with it, And if you didn't fit in, you're lost. And what Pocket Mentor does, what practical AI does, AI is built to think with the human, not for the human. And it goes beyond human in the loop. Human the loop is just a checksum. Practical AI and Pocket Mentor actually works with the human. And AI in general is this great universal translator between the neurotypical and the neurodivergent. So we eliminate barriers to communication. And when it all comes down to it, we look at the best leadership programs that are out there for the last 150 years. They all boil down to that one thing: learning how to communicate, identifying roadblocks and overcoming them. That's what we do here. Pete Vera, Exit Algorithms: Yeah. I love that perspective. Ver very unique. I think people are becoming more educated about that topic and you know, it's not as taboo or or anything. how has your journey been with with Amazon and and is are there any things you you You know, you learned from that experience that have carried over into practical AI now. Derek | HumanFirstAI.net: absolutely. Yeah. Amazon, all the good things and all the bad things that you and your listeners have heard is probably true. And Amazon is so big. Last time I checked, it was one and a half million employees worldwide. And so you really are a small fish in a big pond when you work at Amazon. And my takeaway, my very first manager that I had at Amazon, the one who hired me. she told me this. She said, Hey, while you're here, treat this as if it's your own business. You're a solo entrepreneur, keep track of your wins, maintain, you know, and and absolve your losses. And that was the best advice I had, treating that role as a business. Now, if you're in that whether it be Amazon or Meta or Google. Any of the ones that are they're huge and fast paced like that, you really learn to move at a much faster pace. And when you exit that type of company, you come to realize that you did so much in so short amount of time that when you go to work for somebody else, it's like your mind is thinking so fast you're out thinking everything and you get bored. It's like, why isn't everybody moving as fast as we did? And the reason being is they didn't have all those ducks in a row, so to speak, to to keep that communication flowing. And it really comes down to communication flow and an organization like Amazon, there's just so much going on and so much expectation that we end up living up to our own expectations. or you know, some projects might just blow up and we just accept that. And Jeff Bezos even says, We're okay. With seventy percent of our projects failing, it's it's that the really only about five percent that really make us the big money. So they expect failure. So we fail a lot at Amazon to make those achievements. Pete Vera, Exit Algorithms: That's awesome. Yeah, that's a that's a great cul culture to foster. yeah, very interesting. I you focused on developing teams or or and creating a workplace culture, right? I I'm I'm curious, what do you think is the biggest mistake companies make when they try to train and and develop their their people? Derek | HumanFirstAI.net: Well, I think the biggest mistake is not spending enough time and not giving enough time for that individual to actually learn the ropes and communicate. the the infrastructure that I implemented at Amazon, the engineers, reliability teams, the maintenance teams, the ones that kept the hardware moving and up to code. They had an 83% attrition rate, which means every year for every 100 people they hired, 83 walked out the door because what they used to do is just hire somebody and they'd throw them in the deep end of the pool. when we're instructors and say, Hey, you know, what about this one? You think he's gonna make it? yeah, I think this one's gonna make it. And that, nope. They went under. So yep, hire somebody else. And it gets so expensive. It's expensive on the company. It's expensive emotionally for the individuals that are involved. So the big overlook here is getting that individual onboarded in the correct manner, giving them time. And then once you onboard them, allowing them to make decisions and truly empower them. And it it really was part of that infrastructure that built practical AI and Pocket Mentor because having the right information at the right time makes the individual employee feel worthwhile. And that's that's usually number one on why do you work for a company? It's because I want to do something. I want to be productive. Pete Vera, Exit Algorithms: Do you have you how have you applied that to, you know, building out practical AI? has it affected your you know, obviously your your approach to hiring and selecting people? Derek | HumanFirstAI.net: Yeah, absolutely. The the company itself is it's a voice-based use of AI and it's the interface that's real important here. there's talk of you know, school age kids, students really of all ages, that would copy and paste term papers from ChatGPT and they wouldn't learn anything. And so even when an employee is on the job and they're looking to Fix a problem, either fix a problem that just happened, or learn to build something better, they have to drop everything they're doing, take their hands and their eyes off the prize and turn around and either read a book, read a manual, or some digitized version of a book or a manual, you know, PDF, Word docs, Wiki pages, etc. So they can only do one or the other. So our value that we bring in here is the value of the interface. So you mentioned in the introduction, you know, with a simple phone call, anybody can dial a mentor and have that mentor walk them through whatever the problem or the job or the SOP that they're working on. And when you interface with another human like you and I are doing today, I can focus on you. I don't have to take my hands off the keyboard or mouse. You know, we're here at the computer, we can still click and send messages and and adjust camera angles, and I'm focused. I don't have to look away to reference something. And when somebody's on the job, that's what's most important. Keep them safe, keep their eyes up, keep their hands up, and keep them working. And the in this case, the mentor, whether you're calling, you know, your favorite manager, your dad, your grandfather or grandmother, you just, you know, click that little earpiece that we wear all the time now and say, hey, I'm in a bind. Can you give me a hand? And if if you reached out to me or I reached out to you, that would be the conversation. We just talk each other through it. The difference with AI is that AI allows that knowledge base to be scaled up to the point where we're democratizing information so that every employee across the board is working with the best information right when they need it without stopping and taking a break and going back to the office. Pete Vera, Exit Algorithms: Hmm. I think a lot of businesses could could use that technology or you know, implement that. it sounds like you focus on the so certain industries that that that w works best or are you seeing it at work in other businesses' models as well? Derek | HumanFirstAI.net: Well, knowledge is power, whatever infrastructure, whatever industry that you're in, whether it be blue collar work or white collar work. My background, I came out of industrial trades. I told you, you know, building factories for automobile manufacturers. but that's what really that was the leadoff. Right now, I think every one of your listeners has heard a headline somewhere that college isn't the right path anymore. It's trade school for somebody that's 16, 18, 20 years old, because we have this baby boom generation that is exiting. And the trades have been short on employees for 30 years. And so this was my entry point, my ramp to use AI in a place where AI can't replace the employee because one, there's not enough employees to go around. And it's a way to augment their training. So blue collar work is very much hands-on. they already have radiophones on already, and there's no additional hardware that needs to be purchased. It's just, you know, you tap the button and you talk to your mentor. And it could be how do I change the oil in this gearbox? Why is this oven smoking? You know, how is this HVAC system working? How do I tune it? What's the right pressure? So that's what got me on board. But information, if we need it, we need it. And it could be we're at an office and we have to change the toner once a year and I get lucky to change the toner because it went out on me. How do I know what toner? I could just hit the magic button and say, Hey, what's the toner for this HP laser jet that's here in front of me? And it provides the information. Now I could find that information on my own, but how long would it take? And if it's there instantly, I'm calling my best friend. So that's a simple portion of it. But if you take any SOP and you ask earlier what the major oversight is or what companies overlook when they bring in employees is that they don't spend enough time training them. So with this voice AI that is nestled into this curated box of knowledge, we can teach that AI only what's on the SOP for a job role. And that employee can take that assistant that you we've heard of personal assistants. Pocket Mentor is a professional assistant. It goes around, it's trained on one topic, your role, and you just tap the button and say, hey, I'm in this room. What do I do? Or what time is this? Or what time is that? And it it make sure that employees aren't on their own anymore. And it's a way for companies to continually train their employees with much less money out of their pocket and at the same time allow that employee to be trained while they're actually doing the work. Pete Vera, Exit Algorithms: sounds like there's a lot of applications for it. and we talked a a little bit at the beginning, you know, the pre-show about your book, Human First AI. I'd love to if if you don't mind sharing kind of the the through line of that book, maybe some takeaways business owners can can implement. and yeah, what do you mean by human first AI? Derek | HumanFirstAI.net: Yeah, absolutely. A lot of businesses, a lot of individuals, whether you're an employer or employee, think that there's only two options. We use AI to solve all of our problems, and at the same time we get rid of workers. or you know, AI is just the devil. We need to stay away from it and not use it. And in my book talks about the third option. And human first AI is actually the philosophy behind practical AI and Pocket Mentor. Human first AI means putting the human first. Because I think I reference seven different traits that humans bring to the table that AI can't. The most major trait that AI does bring is the ability to remember on the spot and provide information when there's a knowledge gap. So you take the seven disciplines that humans do best and combine it with that instant knowledge retrieval, that makes that human superhuman. It it truly does empower them. So the book explains that process. it talks about the what ifs we go down path A versus path B. And through the book, the reader learns that AI is is not a villain, that AI can actually be a friend to both business and employee. And as you get through the book, later on in the book, there's a manifesto that explains the walkthrough how every business, whether it starts a f a mom and pop small business all the way up to you know full enterprise at the you know fortune 50 level on how to adopt and to recognize problems. and one thing that we need to understand is that with AI, we shouldn't just throw AI into our company just because we're trying to use AI. We need to truly identify what the problem is. And then solve for that problem. It comes down to Occam's razor. If the solution isn't AI, well then you shouldn't use AI. And a reader reminded me yesterday of of a portion of the book where it talks about when electricity, not invented, but was first adopted and industrialized, you know, let's say what, 125, 150 years ago or so. we had the DC and the AC, but electricity was so new. The headlines back then read like the about electricity, read today like they do about AI. It's good, it's bad, electricity kills people, AI takes our jobs. And they threw electricity at solving all the problems in the world, and some of them stuck and some of them didn't. when I was doing my research, I learned that the first electric. Powered automobile was actually invented in the 1800s because they were trying to find this application for electricity that really wasn't ready for prime time. And we're trying to do the same thing with AI. So those people that fear the ads, go take your job. Well, ninety-five percent, I think that MIT report that came out last year, ninety-five percent of corporate enterprise companies, Pete Vera, Exit Algorithms: Hello. Derek | HumanFirstAI.net: 95% failed at their AI implementation. And that's because they're trying to replace workers instead of working with. So be patient, trust the future. And that's what human first AI does. It's a guidebook to walk any business, small to large, through that adoption to success. Pete Vera, Exit Algorithms: how are you leveraging AI personally? you know, how how do you do you use it for anything you know, besides business? Derek | HumanFirstAI.net: Absolutely. It's it's it's my day-to-day assistant. I use it for knowledge storage, knowledge retrieval. My ADHD combined with my autism and my dyslexia kind of puts me at a situation that even though I'm I'm a certified project manager and I understand how all the categories and and and segmentation should go. When I try to classify something with my own brain, I have a little bit of difficulty keeping track of details and that. So I use AI tools and I'm not really prone to one over another because we are at our early stage infancy. Just like 120 to 150 years ago, people were may have been fighting about, you know, do we use Tesla or do we use Edison base? We don't care about the company. AI is actually a commodity that is it it just serves power. So Chat GPT is great. they improve it. the ChatGPT projects is fantastic. they've added some they call Codecs, and and now ChatGPT and Claude and Perplexity all have these interfaces with your laptop or your computer desktop. So it's there to support you. I use Microsoft Edge, which is a version of Chrome. Microsoft Edge comes with Microsoft Copilot, which is a relabeled Chad GPT variant. And I can go on any page, highlight something, right-click, and have Copilot summarize it, explain it to me, search it. So it augments by day. It helps keep me in alignment. I have some AI that when I put a let's say on my calendar, if I'm traveling somewhere, whether it's to a school appointment or a you know dental visit that if it's from two to three o'clock, my AI will actually read that and see where I'm going and automatically add travel time to block out my calendar so I don't have to do that manually and think. And it's different things like that, the subtleties. We're not we really don't need AI to solve the world's problems. If we just use AI to solve our day-to-day problems that trip us up. That's probably eighty percent of what we do in life. Pete Vera, Exit Algorithms: There seems to be con some convergence too, I think, among the different LLMs. I'm using perplexity, I'm using the browser comment and it like that exact same feature you said with Edge, you can do with with this browser too. but I think you're right, you hit it on the head. It's it's it's great it's a great tool to just optimize your Well, optimize your life if you really use it properly, right? there's there's so many tools out there these days. how does a how would a person or or business owner decide between so many tools? Is do you have a process that you go through to to separate the you know, the wheat from the chaff? Derek | HumanFirstAI.net: In the book Human First AI, we have a walkthrough to do just that. And it's it takes the reader step by step through identifying the problem and finding out what that boils down to. There's a quote that's I find attributed to Albert Einstein. Whether it's a true whether he actually said it or not, I don't know. You know how the internet is. But the quote goes, if I had an hour to solve a problem, I'd I'd spend 55 minutes just understanding what the problem is. So if you understand the problem deep enough, the answer is going to appear. And it could be a specific AI, it could be a version of an AI, it could be old versus new, new versus old. And there's so many solutions out there. the way to actually find out is just to experiment. It truly is. Experiment with what you have. you know, maintain your level of personal safety, you know, when you get in. Don't share anything you're not comfortable with. But just start asking questions, making applications, maybe get a laptop that's old and you don't use it anymore and fire it up and and use it as an air gap to to do just AI experimentation. That way you don't have to w live in fear of it accessing personal information. all these AIs, especially the big companies, OpenAI, Anthropic, et cetera, they have built in security just like any software does. limits its access to your personal data. And that's all AI is, is different software, but that doesn't mean that it's not new to us and anything new. Now we're dealing with change management, which is another leadership, you know, class we're going to talk about Friday from you know 4 to 6 p.m type of thing. So it's it's just understanding that it is that it's here and just just learn what it is. Pete Vera, Exit Algorithms: Yeah. Derek | HumanFirstAI.net: And don't force the solution, just just learn the problem deeper before you make any choices. Pete Vera, Exit Algorithms: have you seen it been deployed most effectively where, you know, they just the business chooses a tool and implements and then all the employees just hop on board and and go for it. is there's some some strategies around change management that you've seen be effective? Derek | HumanFirstAI.net: Well, any change, d to be frank, I've never seen any company and all the employees just jump on board a hundred percent. It it just doesn't happen. We usually teach when we teach change management, I break it down into four quadrants. and it begins with those that you know don't want to adapt no matter what, and it ends with those that would adapt anything because they're just excited for change. And somewhere in the middle is the majority of of humans. And and it's a human decision. It's psychological. It's that human factor. So it really doesn't have to be in the context of AI specifically, but anything that is change. humans for the most part fear change because now we have to it means the future is unknown. So when you bring in any change, whether it's AI or you know different colored uniform, it's important to Communicate with your employee base as far in advance as possible. I mean, 12 months isn't too far for pretty much any change. And then how much you deliver, maybe you sprinkle it in 12 months in advance, nine months in advance, you get a little bit more than six months, you start posting start dates and the enablement dates and how-tos, and you put up your wikis and and your FAQs. And then in that final six months before you adopt something new, whatever that is, you have round table discussion, you invite employees in on the clock, off the clock, you you have flyers out, you post it on your internal news channel and maybe put notes in there. well, people don't get paper checks anymore, but that type of thing. if it's a company newsletter, just communicate, communicate. Communicate and that will break the ice more so than any other thing in the world. There is no magic wand, it's just about communication. Pete Vera, Exit Algorithms: Hmm, yeah, well well said. And Derek, this has been a a great conversation. give you the floor. where can people find you the easiest? Derek | HumanFirstAI.net: Yeah, absolutely. You can I I live in Indiana, so Google me. I win if anybody gets that reference. But you can find me on LinkedIn, humanfirstai.net. I'm happy to give every one of your listeners a free chapter of of the book. don't even need to plug in your email. You can read a chapter right online. Practicalai.app is my work domain. And I'm just happy to answer any questions that you or your listeners might have. And specifically to your domain and what you teach about exit strategy. I've had companies come to me and they say, Hey, we think it would be of value if we can digitally document all of our SOPs. So when we sell a company, the receivership sees more value. because you know, a lot of times companies when the CEO exits, 90% of the knowledge is gone. And when we document that with our practical AI applications, it it makes sure that that knowledge is there to be asked and and have a QA conversation with years after the original CEO exits. Pete Vera, Exit Algorithms: great Derek. It was awesome talking to you. Thanks for coming on the show today. Derek | HumanFirstAI.net: Thank you so much, Pete. good luck and continue your good work here, please.