Subscribe
Share
Share
Embed
Join Derek Hudson as he explores Essential Dynamics, a framework for approaching the challenges facing people and organizations. Consider your Quest!
Welcome to Essential Dynamics. I'm Derek Hudson, host of the Essential Dynamics podcast and creator of the Essential Dynamics Framework. On this podcast, we learn about the principles of Essential Dynamics through deep conversations with interesting people. And I'm very excited to today to have found John Munsell, who is, I think, an interesting person. I am expecting a deep conversation.
Derek:John, welcome, and thanks for joining the podcast.
John:Thanks, Derek. Appreciate you having me on, and and hopefully, I will be interesting as so
Derek:Well, you know, sometimes people say that I ask good questions. And if if you get a good question, you're probably gonna come off as interesting. But I have checked out some of your stuff, and I do think you're interesting. I'm really excited to have you on. And I would love for you to just share a little bit about your background and in particular why you were interested in having this conversation today.
John:Alright. Well, so I spent sixteen years in financial services. And since you're a spreadsheet guy, one of the things that I built back in the early days so I'm I'm old enough to know what the original spreadsheet was, which was VisiCalc.
Derek:VisiCalc. Yep.
John:Then we went to Lotus one two three, and I worked for a bank at the time, and we built a spreadsheet application that analyzed commercial financial statements so we could make lending decisions. And then fast forward, I moved to another bank in Florida, and they didn't have that tool. So I took our tool, and then I made more enhancements to it, built some macros in it. And then I called a friend of mine because he and I both built this thing when we were in this other bank in Baton Rouge. And we said, shoot, let's see if we can't sell this to banks.
John:And so we packaged it up, and then I had to go throw myself into the world of direct mail marketing and direct response copywriting. And so I learned all that, and we started to sell it to banks all over the country, and I learned a lot about marketing. And then in '97, when the web came out, I thought, okay, well, that's the new frontier, and it it tickles a lot of my innate, I guess, interests and and aptitudes. And so I ejected financial services and started a company that built web applications. We gradually became a digital marketing agency that also built websites and web applications.
John:And then this thing called AI came out, and I thought, oh, that's gonna change everything. So I sold off the agency side, and I spent a year working with CEOs to understand the problems that they wanted solved. I was thinking I was gonna be a CMO, and it turned out they were more interested in AI. So instead of being a CMO that knew AI, I decided to really start teaching people AI because that's what they they wanted. And so since I went from a high of 40 employees down to a low of two and a half, I had to use AI to survive and thrive as a business, which is what, you know, led me to develop my own frameworks.
John:And as I started to build those frameworks, people became really interested in those. You know, off we went. And now I'm it's it's been a wild ride, and every every week gets wilder, I should say.
Derek:Well, it's it's fun to be part of so many sort of breakthrough changes. Yeah. And there's patterns, but they they don't necessarily repeat the same way every time. And so because you mentioned spreadsheets and then the web and now AI, that's some something that I'm thinking about. So let me let me put my hypothesis out for you.
Derek:And then and then let's let's see what you can do with it because I am very interested in the in the potential of AI to help people run their businesses better, which I think we have a common interest in. So Yeah. I started my career at Pricewaterhouse as a staff accountant. And in 1984, and we didn't have access to spreadsheets. We did not have access to VisiCalc.
Derek:The clients did the calculations and we checked them. And in fact, we had a contractor who operated a comptometer, and it was Trudy, the comptometer operator. So she'd bring this suitcase in, set it up on a table, then we had to get her a phone book so she could sit on a chair high enough to run this thing, which looked like an old time cash register. Right? All these buttons.
Derek:And she would add up the general ledger to make sure that it added up because we didn't trust the computers back then. So that was pretty low tech. And then eventually, we got our hands on Lowe's one two three. And I was involved in some of the early sort of work at in the whole firm in Canada in terms of how do we use this tool. And for an accountant, it takes away, like, 90% of the drudgery of trying to add those numbers up or change a variable and run it through.
Derek:I've done some extensive financial modeling and it's it's great and I love it. But I'm not sure that it's always been net positive for an organization. And I was thinking about there was a time when we didn't typically or we didn't work remotely. Mhmm. And we didn't even have a computer screen in a boardroom.
Derek:And so the way you communicated your spreadsheet work was by printing it out and bringing stacks of paper into a meeting, which people didn't look at. And I found out that if I had a good summary page and a tall piece of paper, I could probably get anything approved because people didn't I mean, they couldn't pour through the stuff. And and what I've found since is that what should be integrated into an enterprise system in many cases, in many organizations all over the world, is still a one off spreadsheet that might link to another spreadsheet or maybe the link's broken that needs manual updating. And so we don't have that integration. And I don't think that integrated computer system is actually gonna run a business, but we're living way below our potential even in that area.
Derek:So then I see AI come in, particularly with the chat interface, which I love and which is super awesome. We have the potential to optimize all kinds of individual parts of the business and still not solve the problem that our organizations aren't designed for flow, that communication and information value don't flow through organizations. And we're just getting better at producing those thick reports with AI, which at least now we don't have to print. We can just email them to people in a massive PDF. So that's a lot throwing at you there, but I would be interested in anything you wanna pick from that and and put it back to me.
John:Well, so I I I guess what I'm trying to grasp is when you you originally said you had an hypothesis that you wanted to throw out at me, and and I I'm trying to understand wherein that was the hypothesis.
Derek:Yeah. And where's the question? Fair enough. So the hypothesis is that these electronic tools can automate and enhance work that we do in a way that, you know, before you use them, you can't even imagine in in terms of your ability to to understand and manipulate information. But I feel that spreadsheets didn't solve the main problem, which is that organizations aren't organized to for everyone to understand how to do the best work as a as an entire system.
Derek:And I think AI has the potential for sub op like, local optimization of all kinds of job functions.
John:Mhmm.
Derek:And in that way, it would be like a spreadsheet for ideas. And so my hypothesis is if we don't do something different, instead of having a thousand spreadsheets that run an organization, we're gonna have a thousand spreadsheets and a thousand individual AI engagements, and we still won't know how to run our businesses. So that's that's my hypothesis.
John:Gotcha. So I I would say a couple of things. You know, what what I try to focus on is well, let me back up to your point to a degree. The I think the problem that people if they're not aware of it and they don't know how to deal with it, AI just gives you a ton of volume of information that people think, oh, look look how brilliant I am. I just produced a 12 page document in seconds.
John:When a 12 paragraph document would have sufficed, you know. And so now what you're doing is you're shoving a massive amount of information to somebody else who's gonna use AI to digest it into the 12 paragraphs. Right? So, you know, you you end up creating a a bigger mess. And so when you start to read all the research that's out there from all these consulting companies, I think everybody's kinda missing the point, and I think training, and this is where we're heading with our training, training has to kind of catch this before it gets out of hand.
John:And and what I mean is AI will give you this false sense of of security and this false sense that AI is smarter than you. And so you abdicate your own domain expertise to this machine because it's so damn fast and produces massive amounts of volume of information. And so then you end up literally abdicating your expertise to a machine, and you you I think you minimize your usefulness to the organization. But if you are conscious of that as an employer, and you make sure that you're training people how to use more of their domain expertise and not less, and how to see AI as subservient to them rather than their master, you have to teach them how to think differently. And and so what the way I'm looking at it is are are you familiar with book, Corporate Life Cycles, How Businesses Grow and Die and How to Head It Off in the Past, basically?
Derek:No. But thanks for that recommendation, and we will link that in the show notes.
John:Yeah. So fascinating book, very complicated read. I read it fifteen years ago, and I've used it, you know, to scale our business. And my the chairman of our board is the one who said, you gotta read this as you scale. So in in that work now he wrote this, I wanna say, in, like, '99 or '97, somewhere in that range, but but everything he teaches is still true.
John:And and so he uses these these four letters for basically the type of person that you are and the type of role that you play in the organization. You're either a producer, an administrator, an entrepreneur, or an integrator. Okay?
Derek:Mhmm.
John:So the producer is the guy that just likes to do the work, you know, shove it, you know, some food under the table every now and then, and I'll just keep working. You know? The administrator is the person that likes to put rules in place, likes to follow rules, likes to put structure around it. Your accountant is gonna be typically an administrator. The entrepreneur is the idea person.
John:Right? I I don't like the fact that he uses the word entrepreneur, but I guess he didn't have an eye that collided with integrator. So we're renaming all those. We're we're calling the integrator the connector because the integrator is the one who who likes to create community, likes to get the the team spirit going, likes to create culture, etcetera. Alright.
John:Now we're all a mix of all of those things, and and some were a little bit stronger in than others. If you tested me, for instance, I would be very high on the entrepreneur side. I would be extremely low on the administrator side because that stuff drives me crazy. We just finished the two day workshop for Tulane University helping them stand up their AI governance. Governance in me is like baking soda and vinegar.
John:You know? I just I know I have to do it. I am competent at it, but I cannot stand it. And so fortunately, I have somebody on my team who loves that stuff. Okay?
John:So anyway, my my point is is that if you think about what what AI is now AI is gradually taking over those roles. Okay? But but so AI is gonna do the producer role. So it will produce at massive scale. Okay?
John:It will do the administrator role as long as you're giving it what rules to follow. So we will start to to create that structure. Now if it takes those roles and you're up you're mostly a producer, then then what does that do for you? That either replaces you or it it makes you have more free time to explore those other areas, or you're you will be more valuable if you take your domain expertise and drive it through AI. So in other words, you're what you what you can't do is take the same workflows and shove it into AI.
John:You have to be in a position to reimagine the workflows because it it it changes things. It changes a lot of things. And so what I'm trying to to do is teach people how to think differently with AI because it it will mesmerize you. And then because you're reading all this other stuff, you do think that this is why I hate the phrase human in the loop because it makes you feel subservient to the machine. And you should be the one who dominates the machine even though it'll work faster than you, even though it has access to the world's expertise.
John:You're the only one who knows what excellence looks like, and you have to learn how to drive excellence out of AI. And if you if you are able to make that jump, then I don't see how you lose your job. I see how you make the company infinitely more valuable. I think the ones that will lose their jobs are the ones that cower down to AI and and become subservient to the machine to where their boss says, yeah, the machine's better. We can let Derek go now.
John:Right? I don't know whether that totally answers your question, but I'm I'm I'm so afraid that businesses don't see the opportunity to blend AI with the domain expertise of their employees and take they could take a, you know, a a b or a c employee and turn that into an a a plus employee because the a lot of times the reason that that an employee is, you know, not really an a performer is because they they're either not applying themselves or they're just not as fast as somebody else. But this, you know, that you wouldn't have kept them around that didn't have some sort of domain expertise. But with a little bit of training, you can turn them into a a powerful tool, a powerful employee, and somebody that contributes. Right?
Derek:So I really like this, John, and I'm gonna go back to something you said early in this in this part of our conversation. What you're trying to catch before it gets out of hand is abdicating our expertise to machines.
John:Yeah.
Derek:And so if I go back to the spreadsheet example
John:Mhmm.
Derek:I I had this experience many times where someone's working with me would produce something, you know, using a spreadsheet, and they'd bring it to me and they'd say, you know, here's the analysis. And without looking at any other code, I'd flip through and say, well, this doesn't make sense. And and they would go, oh, that's right, you know, and they'd run off and correct it and bring it back.
John:Mhmm.
Derek:And after a few cycles of that, they ask me, how do you do that? And and and there's a few few parts of it. One is, well, I don't I don't have any trust in the numbers that you put in. Mhmm. And not not that I don't trust them, but I I didn't see it being done.
Derek:And so I'm not assuming anything. Whereas when I am crunching the numbers, I think that I'm, you know, following a logical path and, you know, I I don't I'm not aware of my mistakes. And so, you know, one of my suggestions was we'll put it down and then, you know, walk around the block and then pick it up and say, what would Derek see if he looked at this? And, you know, that that kinda works sometimes, and that was one of the things that one of my controllers did as a sort of monthly review exercise. Before I give the financials to Derek and he's gonna find that one thing, I'm gonna use Derek's eyes to see if I can find it.
Derek:So when I when I take that and think about AI, it's yeah. Everyone has an understanding of the problem, you know, or maybe they know they don't understand the problem. And they understand, like you said, what excellent looks like. And if they can step back from the material that's produced and say, that's a good problem statement. That's a good analysis.
Derek:That's a good solution. That's excellent. Whether or not they produced it or AI produced it, the expertise is in play. But when we get I think you use the term mesmerized Mhmm. And we say, wow, look, because if you put this in, that comes out.
Derek:And we can't there's a black box in between, then we surrender some of that judgment either on the problem statement or evaluating the solution. I think that's what one of the things that you're trying to say that you would teach people Yeah. That there there's a tool here, but we're in charge because we know what we want and we know when we've got it. So let's let's be wise and judicious and a little skeptical and make it show its work or whatever. Is that right?
John:Yeah. Yeah. Completely. I mean, I I can give you examples on spreadsheets, and I can give you examples on copy. But let let's take it very simple.
John:So we we have what we call 10 levels of AI mastery. Alright? And the first level is just you ask a question, you get a response, you're blown away. Okay? You get higher up and you get into building agentic workflows, and then you get into managing agents, and then you get into managing people that manage agents.
John:That's the tenth level. Most businesses that we deal with with well, let me rephrase that. Without exception, the business that we deal with, 80% of their employees are operating at level two or below. You have maybe 10% that are operating at level three, and you'll have an outlier that's somewhere between four, five, or six. They're not even aware of the other levels, let alone, you know, on a quest to achieve that level of mastery.
John:And so what ends up happening is they're so caught up in what AI will do and the speed at which it will do it, then they they just look at it and go, that must be good, and then send it on. So they don't consume it, understand it, and say, this is wrong. Mhmm. And that starts with knowing how to how to interact with AI and knowing how to give AI what it needs to do an excellent job. And and that's why, I mean, they they abdicate that to the AI just because they're blown away at the speed and the stuff that it can pull together.
Derek:Mhmm.
John:And so we developed this thing called the AI strategy canvas, which are you familiar with the business model canvas?
Derek:Yes. Yes.
John:Okay. So it's it's it's framed up like the business model canvas, and it parallels it to a degree. Like, on the top right corner of the business model canvas, you have customer segments. In the AI strategy canvas, you have target audience. Alright?
John:So when I'm working with an AI, I'm trying to build something. I'm going through all these blocks to say, what do I need to give it upfront in order to get the best out of it? The the bottom one where you have revenues on the business model canvases, where you have the request. That's where you finally ask it what to do. The left where you have the expenses, that's where you have the rules.
John:You know, what do I have to tell it not to do and what to, you know, kind of respect. Alright. When when we teach people how to use that and they start interacting with AI, they all of a sudden realize, oh, I haven't given it this. I haven't given it that. I haven't given it this.
John:And and then what they produce is subpar. Okay? And then when they read it, before they've gone through the canvas and they understand it, when they read what AI puts out, they're just reading it really fast and like, wow. That that sounds so smart. That that's got to be real.
John:But when I go through it and I show them why this this 14 word sentence is not good, you know, then they start to go, ah, you know, and I said, look, you gotta step back. You are the expert. It's just fast. And all you're going to do is get mediocrity at lightning speed. We we need to train them to think about how do we get excellence at lightning speed, but they need to maintain their confidence in in their knowledge.
John:Does that make sense?
Derek:Yes. It does. And that's that is resonating with this parallel with spreadsheets. You can do a lot of calculations. Yeah.
Derek:And they can ultimately not add a lot of value if you don't understand the parameters of the problem. I and I I ran into this recently where I saw someone else's work, and there was an element of risk and prediction involved, and there were these different sections of a very complicated spreadsheet calculating numbers where the difference between the scenarios was so small and the noise around the the future risk was so big that they didn't mean anything.
John:Really?
Derek:There there was there's no more information provided, but there was variables that you could plug in, and there was lots of cells, and you could do lots of analysis. But at the end of day, you didn't know any more than when you started. And I suspect that that's the same with AI. I'm impressed with numbers. I'm I'm impressed with words.
Derek:There's a lot of words. I read a sentence. I wouldn't have thought of that. So that's value. So you're giving us this discipline.
Derek:Now one of the things that came to me when you talked about this is if you took AI out of the equation, that's still a very good practice. Mhmm. To be very, very clear on what your inputs are, what your rules are, what your outputs are, what the audience needs. And so you're you avoid this mesmerization problem if you follow your rules. But, also, we're gonna get better results even if the AI isn't doesn't blow us away because we've taken the time to focus on what we can focus on because we have automated part of the task.
Derek:Is that fair to say?
John:Yeah. Yeah. And and interestingly enough, like, I taught nine LSU professors how to use AI even even though they have their own curriculum. And one of them was like, I really like this strategy canvas because I'm using it in things other than AI. I'm using it in our organizational structure and stuff like like, cool.
Derek:Okay. Okay.
John:Yeah. So it's making some sense. Okay. The the the way we designed it was to do three things. One was to help businesses and business leaders learn what AI needs.
John:So if you're trying to develop an initiative, you know what departments need to be involved and and how the thing needs to operate so they can they can literally develop an AI strategy on different initiatives they have. The second is to provide a common language so that when you're building something and you need to involve IT and you need to involve HR and you need to involve PR, Those people understand it at the exact same level, and and they can say, oh, you're gonna need me for this block. Right? And so now you have this common language where everybody understands it, and you don't have that that rub between IT and marketing where they hate each other because one thinks the other is an idiot. You're right.
John:And then the last one is the prompt construction itself to where you get the most out of AI. And you don't need to use that every time, but the when you're trying to build a repeatable process or system, you need to go through the canvas so that you make sure that your prompts contain the right information at the right time using the right structure. And and we have a whole structure, ironically enough, that that came to me because I'm such a spreadsheet freak. I wanted to create a prompt structure that had containers and variables in it so that I could take that same prompt, use it for something completely different, and just swap out a couple of variables and execute a different process so I don't have to rewrite the prompt from the ground up.
Derek:Okay. Alright. So I see that we're running out of time, John, very quickly. Yeah. And I'm I'm I'm just the machine my my intelligence, which I hope isn't artificial, is processing this quickly.
John:You're moving like a real human. So I'm
Derek:I'm I'm really I'm really enjoying it. So what I'd like to do is kinda cap this one off. And and my summary of what you're saying is is that we have to use AI to magnify our expertise and not step aside and let what appears to be expertise take take that position because we will not do as well. I I wanna come back in our next episode, and let's go let's go let's get into the CEO's office and talk about things from the other direction. Is that okay?
John:Yeah. That'd be perfect. Let's do it.
Derek:Okay. Okay. So thanks very much, John, for for sitting down for me today. How can people find you? Because I'm sure many will be interested in what you're talking about.
John:Sure. So they can go to our website, which is bazooka.com, bizzuka.com or ingrain.ai. That's the the book I happen to read or write called Ingrain AI, which is available on Amazon. But it explains all of those frameworks to you, by the way. So it is a how to book.
John:It is not a this is why it's important book. It it explains everything to you.
Derek:Okay. Well, thanks very much, and I'm Derek Hudson. You can find me at getunconstrained.com, and until next time, consider your quest.