Why Your AI Strategy Will Fail If You Start With Tools

Speaker A: Foreign. Welcome to the Mason Duke Tech show. And before we jump in, this episode is brought to you by workforce alchemy.com a place where business owners and executives go to uncover profit leaks hidden in your everyday operations. Today's guest is Ben Tasker, a, uh, leader in artificial intelligence education and workforce transformation. Ben has served as an AI program director and professor. He's built applied AI certifications for more than 225,000 learners and speaks widely on how organizations can humanize AI adoption for real business impact. He helps leaders bridge strategy, skills and sustainable change so that their people thrive in the AI era. Ben, super glad to have you. Welcome to the show.

Speaker B: Thanks for having me, Mason. I'm excited.

Speaker A: So you said that AI strategy often fails because leaders focus on tools instead of people. What are the first questions CEOs and executives should ask before adopting AI in their organization, in your opinion?

Speaker B: AI is changing all the time. In fact, it's changing so quickly that within 30 minutes, by the time we're done recording this podcast and by time listeners are done listening to it, AI is going to change. There's going to be a new news article, some new application. So instead of focusing on tools, you really have to focus on the skills. Just because you have AI tools doesn't mean you have an AI strategy. But the skills behind AI, they have a higher shelf life because how you prompt AI, how you can use AI in your workflows, how you can mitigate risk. Those skills training your employees now have a much longer term impact than deciding what tools you need. Because a tool you might get dedicated to might change and might be sold. They might not be in business anymore. So I really like to stress that, uh, organizations, even though you're moving quickly, yes, we are moving through this AI, uh, between times together, and I can explain that a little bit more after. But you really have to sit down and you have to think about what skills do we want to teach our employees? How are we going to teach those skills? And it's really change management. So the next second question is, what's our change management plan? How are we going to build AI out across the organization?

Speaker A: It's, and it's interesting how you talked about how fast things change. I can remember even a year or two ago, maybe 18 months, realistically, probably 18 months ago, I knew people in companies where they were, and most of those were nonprofits, but still they were organizations. And they're like, you're not allowed to use AI. And I'm like, okay, really? I mean that, like, it was like prohibited in the company. Like, oh, we can't have that in there. Wow, what a change just 18 months brings. Now it's just like, like why would you not. But anyway, we'll get into all that stuff, I'm sure. So in your opinion, what does a successful AI transformation look like when it's centered around human skills rather than just technology adoption?

Speaker B: So Mason, you brought up a really good point there. You said human skills. So for listeners that may not know, I really like to emphasize the World Economic, World Economic Forum's skills taxonomy. The reason why I like, it's pretty simple. There's two skills lanes. AI skills, systems thinking, coding, traditional technical skills. Maybe it's prompt engineering. And AI is going to be able to do those skills really well. But right now they're paying a premium because humans are still doing it. And then the other swim lane human skills. Think empathy, think leadership, think communication. Yes, AI can resemble and mimic some of those skills, but those skills are going to be universally human. Humans are still going to excel at those skills. Because just because you can resemble something and mimic it doesn't mean you actually can do it. And humans can do it. AI, no matter how good it gets, won't be able to do it. So human skills are important to think across all roles along with AI skills. And then I kind of like to think of an organization like a cake. So you have the base of the cake, the foundation, that's your frontline workers, maybe that's 80% of your worker base. How are they going to use AI really? That systems based approach, really understanding the skills and the change management and then linking what you want these employees to do to actual learning. So you have to create a learning culture and then you get to the middle of the cake and that's kind of your AI enabled specialist layer. So these are individuals that may already have advanced degrees like myself in data science and AI. They might be robotics engineers, they could be project managers, they could be business leads, program managers. They're not just your front door employees. They're the employees that can work behind the scenes and they're more operational, but they might need a little bit more training, a little bit more depth. They might need to understand how to code a little bit to set up your AI systems where the foundational crew, they might just be doing AI to help. And I hate comparing it to automation, but they're going to be using AI for more advanced automation. And then you have the top of the cake where the frosting and the sprinkles set and that part of the cakes are leaders. So how are the leaders communicating to the rest of the organization why AI is important? How, how are we transitioning for the future? How doing this now actually helps protect your job versus lead to elimination. And it really is more m methodical than just a chainsaw, uh, approach. So by incorporating that change management, by slowly baking the cake, it's really more effective than organizations that are trying to rush through this quickly, implement some tools. And then if you look into this, 95% of organizations that forget the change management or forget the upskilling, forget, forget the reskilling, don't consider the different tiers of their organization. 95% of those organizations fail in AI implementation. So what that really means is they spend a boatload of money and they get nothing out of it, whether the AI is successful or not, just because they don't have the people behind it.

Speaker A: I spent a lot of time in HR strategy sessions and we've talked about with, I've talked with senior leaders about how do you adjust in strategy and what matters and what's important. And in the old days, it used to be a little easier. You had these. If you're hiring someone for accounting, they had accounting skills. If you're hiring someone for engineering, they had engineering skills. And what a lot of them are talking to me about now is adaptability. It's like, really what you did before was great for the way the world worked before, but now the things are different. It's not what do you know, it's what can you learn, how quick, how can you adapt and assimilate and grow. And I'm hearing people talking about things like coachability and adaptability way more than, oh, well, they got straight A's in their computer science classes. That's great.

Speaker B: You're spot, you're spot on, Mason. Adaptability and flexibility are, uh, two of the top skills. They actually outrank all of the AI skills. And I think it's going to continue that way just because organizations are rapidly changing and you need those two skills to change in these times.

Speaker A: And I think about people who, I know you talked earlier about how, how fast things are changing. Like, I just learned to use this system and I really mastered it this way. And all of a sudden that system's gone, something's new. And then you see them all frustrated, like they just wasted their time and like, uh, but no, like, you got to keep evolving with it. But you'll get me on a rant there. But I, I, I'm fascinated by this topic which is why I wanted to invite you on. So how should executives, in your opinion, balance long term AI ambition with real world risks, such as, let me see, ethical concerns and data governance issues.

Speaker B: So all AI, even more simplistic AI, All AI has risks. I like to call it responsible AI. So it's not just compliance. Responsible AI takes individuals across the organization and you really weigh out the risks of whatever you're trying to implement and you do that before any type of implementation. Yes, it might take a little bit longer to do this, but if I'm trying to implement even a simple chatbot to answer some IT tickets, you should have some people from IT there, you should have some people from HR there. Just because what happens if you misclassify or whatever? You want to have some business owners there, people that actually understand the ticketing system and how that works today. So there might be 10 to 15 people on this, but then that really helps set up that system. You can then create some guardrails and some rules like what happens if we do misclassify a person or a ticket. Or what if, you know, if this is more agentic AI, which means it's more automated, what if this tool gets a little bit more intelligent and does something it's not supposed to do and damages the system because it thinks it's supposed to do that? How do we mitigate that risk? How quickly can we shut this off? Is that baked into this or have we not thought of that? How are we measuring the prompts that are going into it and the user engagement? Are we doing that? Is it anonymized or de. Anonymized, meaning that we can look it up and see what everyone's doing, or is it anonymous where it's random, but we can still see what everyone's doing. And depending upon that recipe and the different levels of risk, then you can really set up a, uh, successful AI implementation. But I think it's really important to understand that in this responsible AI, in this responsible AI roadmap, you're also putting into the change management. So I just want to re emphasize that this is part of that change management. I'm not calling it risk and compliance. It's its own separate entity. There's, there's going to be bias risks, there's going to be accountability risks, like if this thing does hallucinate, who's responsible for it. And yes, some things are lower risk, but a lot of people want to go for the highest level of AI without understanding the lower levels of AI, and companies are going to keep doing that and it's just going to keep getting riskier and um, riskier.

Speaker A: So you've helped build programs that made thousands of professionals AI ready. What are some common misconceptions about AI skills that most leaders still have?

Speaker B: It's fast. You can do this really quickly to learn new skills at the rate you need to learn new things, learn them. Especially if your organization hasn't really explored AI. To your point, if they shut it off and now we're just getting into it takes about 18 months to bring someone from uh, a level one, they don't really know what AI stands for, let alone use it, to a level three where they can successfully prompt a chatbot, use uh, it for common tasks, take the output, iterate on it, put it back into the chatbot. That takes 18 months to get someone to a 4 or a 5. Takes even longer than that on a capability scale. In some of these AI domains you need advanced degrees like robotics, for example, drones that might take five to six years to train. And a lot of the executives are like, well, how quickly can we implement this? My, my answer to that is how quickly can we upskill and reskill our employees? Even if these aren't skills that we're looking for now, all jobs are going to be impacted by artificial intelligence. We might as well start with the reskilling and the upskilling because of that change. But also, even though on paper this might take time, there's cost to that. It's cheaper than, less expensive than just going out to the market and hiring for each niche, niche skill that we need. Because you're never going to need.

Speaker A: I was going to ask about that. I was going to ask about that. Like do you have any specific tips on uh, in that area between the upskilling and hiring? What are your thoughts there?

Speaker B: I mean I think upskilling is extremely important. You're going to have to develop that for all your employees. I think for super niche projects, maybe robotics, drones, it's really dependent upon what your organization does. But if it's super niche that you might need a specialist for, then that's when I. Because you can't really train for that. Right? You have to have that expertise. With that being said though, a lot of organizations when they think of AI for some reason think of the more advanced versions of it, but they don't, can't see what's directly in front of them. So I also recommend maybe focusing on some of the lower hanging fruit for the roi. So what processes and problems can we fix? Is there a form that individuals need to fill out that you can just create an agent for it, fills in the form and then submits it. Maybe timesheeting can be replaced with an agent. Kodak has a AI agent with their driverless truck system. And the agent asks each driver before they, before the door unlocks on the truck whether it's driverless or not, to understand if they're feeling well that day. And if they're not feeling well that day, they're asked to go home and they're still paid. But it's, it's a safety concern. So how is that something that your organization can do? 8 questions. The email, the manager's notified, they can't get in the truck, they're not losing pay, so they're incentivized to learn. And that when I say change management, that's what those are the things that I'm talking about.

Speaker A: So you talk about human centric AI. What, um, does that look like in everyday business operations?

Speaker B: So human centered AI, exactly what it sounds like. There's a human in the loop. So all these decisions just aren't made in a boardroom. You have the entire organization experimenting and using AI, finding pain points where it may be successful and where it may fail. And that learning, right, that's creating a learning culture. That's another way you can do upskilling and reskilling because you're sanctioning it at that point, instead of saying we're not doing it, everybody's encouraged to do it. You can then create what I like to call a backlog or a problem list of where AI might be helpful. And then that can go to the specialist and the leaders. And then you can kind of assign value to some of those complexities. And then based upon that value, you can decide which AI projects and implementations you want to pursue for the year versus the ones maybe you don't want to. And you can attach risk to it better. So even though it seems unorganized because we're allowing everybody to do AI at the same time, it doesn't necessarily have to be. It can be, it can be organized chaos instead of unorganized chaos. But, uh, that's one way that I would approach it.

Speaker A: So can you maybe share a story where AI adoption increase trust and performance rather than resistance inside an organization?

Speaker B: Yeah, absolutely. So I'm going to go back to Kodak. So back in 2018, before driverless technology, physical AI, deep learning technology was even being discussed, they knew that. And Kodak's a camera company. Uh, Kodak has trucks on the road for shipping. And there's going to be less truck drivers in the future because of the demand. It's, they take time to train and there's just not enough truck drivers between that one organization and all other organizations. So they were leaning into driverless truck technology. So think of a semi truck. These semi trucks today can now drive themselves. But they didn't start that way. So initially they had to figure out how to design a driverless truck. So that was part of the change management. Then they had to figure out which routes a uh, driverless truck could go on. Then they had to figure out how the driver could drive in the truck still so that the AI can learn. And were some of the assumptions correct? Like for example, can it park itself? Does it doesn't know how far it can go? Some of those questions that on paper might make sense, but on the road may not. Where and where can it drive? Does each state allow this? Is it only in specific regions? What's the responsible AI? And then how do we get the buy in across the organization? So they really attach to that safety aspect. A lot of truck drivers sometimes do get injured or they have to sit for a long time. It could cause mental illness as well. So and they still got paid while they were training it. So they got paid overtime. So that's a human in the loop approach because there's still trucks at Kodak that don't drive themselves. They have more than 100 truck fleet. It's a very small percentage of their fleet that is driverless. And it's doing very specific tasks that, that a driver probably wouldn't want to do. So they actually need more drivers than less because of that demand curve. So jobs weren't really impacted and they really thought through this implementation system more than just the AI. It was a human first approach, not an AI first approach.

Speaker A: So for leaders that are hiring new talent today, which AI competencies matter most and which ones are just simply hype, in your opinion?

Speaker B: The AI competencies that matter most would be prompt engineering. Because all types of AI that most employees are going to use need some sort of prompt framework or prompt engineering to it. And it might sound simple, but there is some art to it. And different technologies use different prompting styles. So you need to know when to use what when you need to know responsible AI. So I would make sure that we're hiring for that. Making sure people understand the risk trade offs, flexibility, adaptability. Like you were saying, that's harder to put in a, a job spec, but it's something that you could probably flesh out in the interview. And I would also try to look for AI skills across all domains because then that can help bake into your upskilling plan. So if you're in digital marketing, for example, and you typically don't require people to know generative AI tools, maybe you put a job spec out there that does require that with all the other specs, just to see, to see what kind of talent you can attract. And then that individual can help, uh, upskill the rest of that department, while at the same time, you know, they're prepared for the future of work. So you're moving together more in tandem. Um, so it's also a little bit of thought into that, right? Like it's not just like, we're not exactly sure all the skills, but we know some things are going to stick around, so why don't we try to hire for some of those skills now?

Speaker A: So how can managers support employees who feel, I would argue, in many cases, justifiably anxious about AI replacing their roles?

Speaker B: So there was actually a recent study in the Wall Street Journal, there's, to your point, more animosity towards AI than positivity. I think that's because the news, there's all these wordings around layoffs and change. And I don't, I mean, there is some positive news, but it's like 100 to 1 in that part's opinion. I haven't looked, but there's way less positive news. But I think AI can have a positive impact on organizations. Upskilling benefits you. It's paying a premium right now. AI individuals and organizations that are embracing AI and that have AI skills are producing a 52% more revenue than organizations that aren't. How that trickles down to the employees that they can make more with these AI skills also by learning them. Whichever way the organization goes, you have these skills. And it's easier to move individuals into new roles that don't really exist yet, but still have AI components than individuals that don't have those skills. It's harder to organize that into the org chart. And this technology is just going to keep changing. AI agents are very popular. There's physical AI, the technology that can drive itself. Who's managing that? Are we going to have individuals that don't know AI managing that? Probably not, right? So even though we might not have it in our organization today, how are we going to plan for that six to nine months from now?

Speaker A: So what advice do you give to professionals who do not consider themselves technical but want to thrive in uh, an AI enhanced workplace.

Speaker B: So AI doesn't have to be complex to learn. Yes, there are Experts that have PhDs, master level degrees, but learning can be fun. You don't necessarily need that level of knowledge to use it for basic knowledge. Jobs like project management, finance, even executive leaders don't need that level of knowledge. They just need to know the skills to apply to the AI technology that's relevant at the organization today. So I would try to make it fun for me. When I started using large language models, AI assisted tools, I had it make a workout plan for me. I took a picture of all my equipment, then loaded that equipment into the AI agent and I asked it for a specific type of workout. I do CrossFit. So I asked it for a CrossFit workout and it was able to spit it, uh, out. I didn't have to type, I had the pictures over time because AI is personalized and knew what we were doing so I didn't have to prompt it as heavily. I then was like, hey, you know, this is kind of fun. What if I take a picture of the food in my refrigerator and ask it to make a meal plan for me Again, all pictures, no words, something I wouldn't do normally. And now it's taking the workout with what we have for, for dinner and it's optimizing for it. A fun way to understand, uh, the limits of the technology to see if it's true. Some of the recipes were not that good. But it tried, you know, it tried. Sometimes it can't determine what you're taking a picture of because either the label's generic or. But that's a limit, right? Like that applies to work. But then how can I apply this fun learning to work? So that's prompt engineering risk, what to put in it, what not like I put in food and workouts. I didn't put in, you know, top secret information in there, right? Like there's. So it minimize that risk and then that really can set up an individual for success. So and that got me comfortable with it. So I could trust it because I think trust is an important factor there. Because if it's wrong on the first or second use, you're probably not going to use it again. You're just going to shut off, especially if you're resistant. So that built the trust and then I could take on, uh, can I use this for coding? Yeah, obviously you can. But is it right? Because we're not sure that it was okay, you know, and this was a long time ago and now today it's much better than it was then. Can I use this for data visualization? Can I use this to make infographics and slides? So I guess the other counterpoint to doing something fun is maybe try to do something or uh, find something you don't like to do and how you can use AI to, to make that task more quickly. So an example of that might be using AI to organize your email box so that higher relevant emails are on top and lower relevant emails are on bottom, just so you can shift through your day. And then maybe you make a task list or automatically update your calendar. That right there can save you up to an hour at least. Now you have an hour back.

Speaker A: So as AI continues to evolve, what are the most underappreciated shifts that organizations are not preparing for today?

Speaker B: I think a lot of organizations, and rightfully so, uh, it's most of the hype, but AI is moving in their mind quicker than it actually is. I mean, AI is moving fast, but I think a lot of organizations already think we're out of that AI between times. So we're still trying to figure out what this AI thing is. And companies already think it's solidified and it's defined and that's causing a lot of mistakes. But I think those mistakes are going to become more amplified. I call them AI whoopsies. Like we implemented a chatbot and it told all our customers something it wasn't supposed to like whoops. Like that's going to keep picking up. So really understand that there's risks associated with the AI that you don't have to move as quickly as every other organization. Doing this methodically actually might pay dividends. You need an upskilling and reskilling plan for your organization, especially as you get to the more advanced types of AI. How are you going to factor that in? What kind of jobs may you need even if you don't have a definition of it? And then how are you going to change the organization's culture? So the change management, this isn't just a regular technology or a ticketing system or a, um, marketing system. It's a technology that can learn, that can become personalized. It could really impact the business. It's much different and it's going to be, it's going to continually change to be much more different. So how are we becoming different with that?

Speaker A: I'm interested in your opinion on this because I know people soft shoe around it and you can see it a mile away. Business there, people that are legitimately scared. My job's going to be gone and they hear from management. I'm m using air quotes for those that are listening and not watching the video that oh no, no, this isn't going to, this isn't a risk to your job. This is about enhancing who you are, which I get, I, I But when the employees who are fearful of losing their jobs hear someone say oh no, no, this is just about enhancing you, and they don't believe them. Some of them justifiably so, not all of course. What do you say for your management? I mean, because that's a tough call and I know I'm um, kind of, that's a little bit put you on your spot. So uh, in your opinion, you run an organization and you got people that some people's jobs are more at risk than others because things are changing for obvious reasons. I get it, I understand it all. But if people are that fearful and you're in charge of the company, what do you say to them?

Speaker B: I mean in my opinion, I guess it depends on how large this audience is. But some organizations have already tackled this problem. So I'm a data driven individual. But uh, for example, Walmart rolled out AI to all their employees and encouraged them to use it. So they gave everybody free licenses. They can use it at home, they can use it at work. It uh, didn't much matter as long as you were using it. And then that then change the culture to become a learning culture and then not Walmart. But other organizations have also embraced if you learn AI, if you um, go on to these learning tracks, if we do try to, to enhance it, there won't be any AI related layoffs for three years, four years, five years. So it's, it makes employees feel safe around that change. Honestly, to get any of the ROI successfully anyways, you probably need that amount of time because if you make a decision too quickly. There's a lot of organizations that think AI is fully implemented. I mean go look up a news story, there's a bunch of them, an entire 30,000 person organization will go away and then the next week they have to hire everybody back because the AI wasn't as good. So yes, artificial intelligence is intelligent. I'm um, using air quotes there for people listening, not watching. But it's not uh, it's not as intense as a lot of people think. And you really need to understand your business. So, so there's a, there's some foresight there, but in my opinion, I think leaders should be transparent. Jobs are going to change. The upskilling and the reskilling prepares workers for those jobs. And it's, it's not a replacement technique. It's when these things do change, we're ready to, to change with it.

Speaker A: I, uh, love what you're saying there. That's kind of what I hoped you would say because you're, you're, you're being honest with them and saying that the environment is changing and if you refuse to change and you refuse to adapt and you refuse to grow, refuse to learn new things, whether it's AI or anything for that matter, that's a choice you as an employee are making and, uh, probably not a good one. Whether it's AI or anything in this particular context, in this particular discussion, we're talking about AI but if you're telling me as an employee, I don't want to learn, I don't want to grow, I don't want to adapt, and I don't want to lose my job, good luck with that one. Whatever your field is in this case, I mean, uh, I mean, who in any. Who in any role regards to AI Maybe it's continual learning in the legal field or continual, uh, what do they call it? Uh, yeah, continual learning in the, in HR or in sales or in marketing or like, like, oh, I, I don't want to learn this new stuff and I want to maintain job security. I mean, come on, that's not realistic anywhere.

Speaker B: So I mean, yes, absolutely.

Speaker A: I like that you, you chose to be open and transparent from m. The managers. And I'm using this in air quotes for people that are just listening. Managers sitting, not, not filled them full of nonsense, but telling them the truth, like, hey, times are changing. We're adapting to keep up and we want to encourage you to do the same. If you're worried about losing your job then, or having your role eliminated, then stay the same. Don't adapt with it, don't improve, stop developing your skills. But if you do that in any role in any company, you've reached your career cul de sac anyway.

Speaker B: And I think that's extremely important to have those learning tracks. And I know a lot of leaders don't think in that sense, but if you have some learning pathways for employees to follow so that they know where to go on the menu, then it's easier to engage with the organization and encourage it. So then it's not just a blanket statements like, hey, I really do believe this. Here's some learning, learning tracks. If you're not seeing something there, let us know. But we developed X amount of them. We're hoping that you, you embark on your journey.

Speaker A: So if there was only one piece of advice that you could give to business leaders who are listening or watching today, what would be the most important piece of advice you could give and why?

Speaker B: I really think a lot of individuals don't consider AI implementation change management. I know I've said that a lot today, but I would really challenge everybody to think of AI anything as change management because there's a lot of dynamics as we discussed today around AI and it's only going to exacerbate. I think there's going to be less positivity around it, there's going to be more negative news around it. Even if AI has a major breakthrough and discovers something that, that people thought was never possible, I think there's still going to be some negativity around it. So I would really think about in change management, how are we going to message this across the organization? How are we going to design those learning plans? How are we going to embrace some of these other changes that come along with the AI and uh, is our organization ready to do that? And if there's more no's and yeses then, then how can you move those nos to yeses so, so that you become ready but just rushing into it because everybody else is, that's not a recipe for success. You're just going to end up as the other 95% of organizations that huh, have tried to implement this and have failed.

Speaker A: So for people who want to know more about you, your work that you do, what are the best ways for them to learn more and connect with you?

Speaker B: You can connect with me on LinkedIn, Ben Tasker and you can go to my website, Fantasy. There's uh, a meeting link there, we can set up a conversation. But uh, happy to help you or your organization discover your learning tracks and your AI capabilities.

Speaker A: Really appreciate your time. Thank you for sharing your wisdom and your insight, your experience. I hope everyone else learned as much and enjoyed it as much as I did. Thank you so much Ben.

Speaker B: Thanks for having me. Mason.