Top Tips from John Munsell:

1. Use Multiple Paid AI Models for Efficiency

“Don’t get enamoured with the idea that you’re going to get it all in one place. In other words, you should have in your arsenal multiple LLMs (Large Language Models). I would probably have ChatGPT, Claude, and either Gemini or Perplexity. These three combined will cost you about $60 a month. But this combination can save you 8-10 hours a week. Make sure you have paid versions of these tools, not the free ones.”

2. Structure Your Prompts and Create Reusable Knowledge-Based Documents

“Quickly learn how to structure prompts in a methodical fashion. Learn how to save those prompts so that you can use them later. Once you produce an artifact like a report or research, you can store it and move up another level, creating a knowledge-based document in a project. This allows you to have system instructions in place, making your processes repeatable and scalable.”

3. Leverage AI to Uncover Insights You Might Miss

“Use AI to uncover things that would normally escape your attention. For example, when you’re in meetings, use AI to analyze transcripts and identify nuances that might have been overlooked, like partnership opportunities or potential concerns that didn’t come up. AI can surface critical insights that help you make better decisions and take action on things you might have missed.”

SUMMARY KEYWORDS 

AI strategy, scalable prompt engineering, AI adoption, AI tools, AI efficiency, AI mastery, AI implementation, AI training, AI frameworks, AI applications, AI integration, AI security, AI productivity, AI customization, AI prompts. 

SPEAKERS 

Debra Chantry-Taylor, John Munsell 

John Munsell  00:00 

Emailing me a copy of the PDF, creating a database of all the responses, creating a dashboard with heat maps and all that stuff that would have taken six weeks. I built that entire thing with AI in three days, and it was unbelievable. You want to customise your own prompt, and one of the things that we always add to ours is I want you to identify the things that would have escaped my attention that will add value or lead to something else in this conversation. 

Debra Chantry-Taylor  00:35 

Welcome to another episode of Better Business, Better Life. I’m your host, Debra Chantry-Taylor, and I’m passionate about helping business owners lead a better life by creating a better business. On the show, I bring on people who are familiar with the EOS tools, and they share with us their experiences. But today’s guest is really special. He is the adjunct instructor of AI at Louisiana State University. He has used EOS in his digital agencies that understands the concepts of processes, core processes, and he is the creator of the AI strategy canvas and also scalable prompt engineering. What he’s going to share with you today is how to use AI to increase your revenue, accelerate your growth, and save you and your team bucket loads of time through using AI to improve all of your business core processes. John Munsell is the author of ingrain and is also the CEO of Bizzuka. Welcome to the show, John. It’s so nice to see you again. How are you? 

John Munsell  01:35 

I’m great. Debra, nice to see you again. I love the EOS symbols in the background. There. We got the framework 

Debra Chantry-Taylor  01:43 

Very passionate about Eos, as you will know. But today we’re talking about something very, very different, aren’t we? So today we’re going to be talking about what you are really good at, which is the AI side of business. So John, for the people who haven’t, haven’t come across you before, why don’t you give us a bit of background about your your history, how you got to where you are now, and what you’re doing now. 

John Munsell  02:04 

Yeah, so I started, I spent 17 years in financial services, and in that role, I learned a little bit about direct mail marketing and financial analysis. And then when the web came out in 97 I thought, Oh, this, this is where the future is headed. So I ejected started a company that built web applications. Gradually got into building mobile applications, and then we built websites, and then we did digital marketing and became a digital marketing agency. And then about four years ago, I sold off the agency side, because my wife and I were about to become empty nesters, Famous last words for those of you who are threatening to become empty nesters. And so I thought, well, this would be a great time to be a little bit more mobile. And so we sold that part of the agency, or like the company, off in that I started to work with CEOs every Friday and teach them a little bit about marketing. And going from a high of 40 employees down to a two and a half, I had to do certain things that I no longer had a staff to do, and so I was using AI to do that. And then, as I was working with these CEOs, they became way more interested in how I was using AI than anything about marketing. And so gradually, you know, I had to develop frameworks. I had to develop processes being an x, e, o, er, e, o, s shop, you know, I do a lot about structuring our company so that we could scale and grow. And so I developed these frameworks or proven processes, as you would say. And I, next thing, you know, I’m teaching CEOs how to actually use AI in a more scalable fashion. So I developed a framework called the AI strategy canvas, which teaches businesses and people what an AI LLM needs in order to give the best output. So it organises it so you can understand it very simply. And then I created this other process that we call scalable prompt engineering, which just allows you to create a prompt when you need to share that process with other people in the organisation. It’s a very structured methodology. And then the next thing I know, the provost of LSU, or Louisiana State University, heard what I was doing. He called me up. Wanted to come over at the time. Debra, honestly, I didn’t know what a provost was, so he came over, introduces himself. I’m the Executive Vice President, Provost of LSU. Do you know what that is? And I said, honestly, no, I. I thought, you know, I thought he was an intern or something, you know, just basically doing something for LSU. And I’m like, Yeah, help you out, buddy. Turns out there’s the president of the university, and then there’s him, he’s, he’s over all of academia across all of LSU campuses across the state. I’m like, oh, I should have worn a nicer shirt, right? So then he asked me if I wanted to teach the honours class the next semester, and I was like, I’m just not really cut out to be a professor, and it was an hour drive from where we live, you know. But I said, I think we could have a bigger impact if we went through LSU continuing education programme and went directly to businesses, and we helped businesses and their employees understand how to use AI so it can amplify their productivity. I said, you know, the problem with universities is they teach people how to build llms, but they don’t teach people how to use them. And the vast majority of us are going to use them. We’re not going to build them. It’s kind of like trying to teach students how to build Excel rather than use Excel. Yeah, the rest of the planet is going to use it. There are only a handful that are going to build it, you know. So anyway, that’s how it got started. And the next thing you know, we started to grow. And we’re training large corporations and small corporations, you know, anywhere from one or two man bands to a company that has 50,000 employees, you know. So it’s all over the board. Wow. 

Debra Chantry-Taylor  06:41 

And you’ve, you’ve written a book that talks about all these things we’re talking about here now, right? 

John Munsell  06:45 

What’s interesting now that, you know, I think about it. We started a group. So I wrote a book called ingrain AI that I released back in March or April of 2024, 525. Sorry. What year is this? I’ve already forgotten, but it came out, and then that was, you know, my my process. There was, I taught everybody how to do this. So it wasn’t a, you know, one of these books that you read that it that tells you why AI is important. This is a, this is a literal manual with frameworks and everything else, very much like traction, if you will. 

Debra Chantry-Taylor  07:29 

The how to book. I call it, yeah, 

John Munsell  07:31 

exactly right. And so if a business wanted to, they could read the book and they would learn how to implement AI in their organisation. But on the other end there, as you know, there are a lot of businesses that say, can you just come and do this for us? And so in October of last year, I launched the ingrain AI certified implementer programme. And ironically enough, Debra, we have the majority. Well, let me rephrase that. The country with the highest number of people in our programme outside of the US is Australia. 

Debra Chantry-Taylor  08:11 

That’s fascinating. 

John Munsell  08:14 

Yeah, so too, yeah. We have like five people in the programme from Australia, which is really cool. 

Debra Chantry-Taylor  08:19 

That is cool, I tell you what it’s certainly, I’m a member of an entrepreneur’s organisation over here, and I mean, it’s definitely people are using AI a lot in the entrepreneur’s organisation, in terms of using it in their business, yeah, but really more around the basic stuff around, you know, helping write things, helping ask questions. But it’s a lot more powerful than that, isn’t it, way more? 

John Munsell  08:40 

Yeah, and that’s the interesting thing, is that most companies that we talk to, they typically have started in marketing because they see the content creation aspects of it. But what will happen is the CEO will become interested in that, because they see somebody in marketing, producing things, and then they start reading about it, and then they get fearful, they get excited, and then they get fearful. And that’s typically when we get called, where the CEO will call us and say, I I know I need to do this, but I don’t know enough about it to lead the organisation. Tell me what else it can do. And so we have a workshop that we’ll put on for their organisation so they understand how they can use AI and HR, how they can use AI in finance, how they can use it in sales, and, of course, in marketing and other departments. We actually work with one of the world’s largest furniture manufacturing for business furniture, you know, office furniture, and they have space planners in there. Okay, I’m not in that world, but I used to be in the office based business, so I knew what space planning was. And so. So we’re teaching them how to use AI in that realm, which is way different from anything I would think. You know, it’s not your typical use of AI. Like I said, we’re not writing blog posts in that space, planning thing. And one of our students, one of the one of the things in our training is what we call a capstone project. You have to produce an AI tool that shaves three hours or more per week off of your job. And so we work with them to produce this capstone project. And one of the guys from this organisation built something that absolutely blew my mind. In fact, it’s so complicated I can’t even explain it. Who you but he has all of these architectural plans, and in those plans, it calls for these desks and these chairs and all these different offices. And so they go out and they order it from different manufacturers. And then those, those documents come in as that stuff comes in. So now he has to match it all up to make sure that all those orders came in. And then he has to produce an artefact that says, okay, all of these, these chairs, go here, these, these desks go, there are these, etc. He built it all, and it’s a massive undertaking. And I was like, Wow, I’m I’m surprised, but you know, it was cool to watch him, him do it. And then he literally shot me a text two days ago, and he said, So Claude just released Opus 4.6 not too long ago at that at the time of us recording this. And he said he he retested his prompts in Opus 4.6 and he was blown away. He said, I can’t wait to show you it was unlike sighted and see what what he does. And then, of course, in finance, you know, I don’t know whether you’ve ever used chat or whether you’ve ever used Claude to build a spreadsheet? 

Debra Chantry-Taylor  12:04 

I personally have it, but I’ve certainly heard of it and seen people do it here. 

John Munsell  12:07 

It’s stunning. And then there’s Claude co work, which is amazing. There’s grok, which does really interesting things. I’m talking about tools outside of just Chantry PT that you can use in your business. But one of the interesting tools that you know, I don’t really ever play with, is one called Gen spark. Are you familiar with genspark? I’m not, no, oh, man, it’s powerful as well. It does a whole bunch of things. It’s like this swiss army knife of AI tools. It’s really fascinating. You do ask, 

Debra Chantry-Taylor  12:42 

You do ask, you’re going to ruin my weekend tonight, because I’m going to spend the rest of the weekend researching. 

12:46 

You’re going to go down some rabbit holes now, absolutely. 

Debra Chantry-Taylor  12:50 

But what I’m hearing you say, and I’m going to type back to Eos, because we obviously talk about the proven processes, the internal proven processes for all of the main functions in the business. What I’ve just heard you say is pretty much every main function can somehow utilise AI to improve their efficiency and shave time off the workload. Totally. 

John Munsell  13:08 

Yeah, completely. And so here’s our philosophy. So you know, if you read or research, you’ll see a lot of research that says 90 some odd percent of businesses have adopted AI, right? The problem is that if you read the footnotes at the bottom of that research, you’ll then see how they define AI adoption. And AI adoption in the research is using one tool in your organisation that happens to have AI in it. That could be a customer facing chat bot on your website. It could be you have Microsoft copilot, just because you have Microsoft 365 it could be any number of things, but it’s one application in the organisation that has AI somehow associated with it. It could be Photoshop. Photoshop has AI in it, right? To me, that’s not AI adoption. That’s just dipping your toe in the water. AI adoption is when everyone in your organisation who uses a computer more than 30% of the day can use AI with their own fingertips at their own desktops to build their own tools to accelerate their productivity, and they know it enough to replicate their expertise inside of AI. So we have this thing that we call the 10 levels of AI mastery, and most businesses, their employees are operating at level two or level three. And I mean, we’ve done this research over and over. Again, and we know that if we can get the majority of employees operating at levels five or six, then the impact to the organisation is dramatic, because now you have literally everybody in every every facet of the organisation, building their own tools, and they’ve learned how to get AI to to produce with excellence. And what I say is, if you don’t know what excellence looks like in your own domain, you will always get average or less out of AI. And so we teach people how to tap into their own expertise and drive that out of AI. That’s where the real impact is. 

Debra Chantry-Taylor  15:46 

I love it. That’s fantastic. So I think you said earlier, though, that what tends to be people get excited by AI, then they get scared of it, and the fear kind of comes in. And what you’re talking about when you’re starting to really incorporate into the business like that, I’m guessing the fears people have is around security, around losing jobs. I don’t know what are the fears people have around AI in the business. 

John Munsell  16:07 

All of the above, all right, so the so the leaders are afraid that the people underneath them know more about it than they do, and so they don’t know how to lead. Now, that’s a legit fear like I have. I have CEOs call me and say, can you train me? And they don’t want to teach anybody until they know it themselves, and that makes total sense, right? And so we will, we will train them. The employees, on the other hand, are afraid that if they learn it and they use it, they will lose their jobs, right? They’re suspicious that you’re just trying to get me to use it so that you can replace me. And a lot of people on the IT side are afraid for the security side of things, which they should be. The more powerful AI gets, the more AI or the more security falls behind, because AI is progressing far faster than security can keep up with it, and so that’s a legitimate fear, but what you wouldn’t do is take your three year old who just learned how to ride a tricycle and slap them on a Harley and say, go at it. You know, you wouldn’t do that. And what happens in an organisation is they, they either know or don’t know that their employees are using AI. If they don’t know, their employees are using AI. We call that shadow AI. So they’re kind of bringing their own AI to work, right? They’re using their own chat, TPT licences, or Claude, or whatever the tool is, and they’re typically using a free licence. The Free licence, then doesn’t allow you to tell the AI don’t use what I upload and train the model, so anything they upload could be distributed inside of AI in ways that you don’t want you know. So we have to teach people how to use it, but we also have to control the licencing so that we can at least protect all of our IP, all of our protected information and all that stuff, so that that part needs to be addressed. But once you teach people how to use it, and you teach them how to take their own capabilities and amplify them using AI, they start to enjoy their jobs more, because they get the all the stuff that is, the drudgery, you know, the things that drive them crazy, the gnats flying around their ear, as I call it. They can hand that off to AI, but they also know what that should look like coming back, right? And so what we try to do is teach them how to do that at scale, so you’re not just mesmerised because you threw something in chat, GPT, and it gave you this voluminous report, and you’re like, Oh, this is amazing. But really, when you dig in, you realise, man, it’s missing a lot, you know. But once you learn how to do it the right way, then you can teach the person next to you, right? And that person next to you goes, What did you do? Did you just do? And, and you’re showing them this report that used to take me two and a half days to produce. I can produce it now in 20 minutes. And that’s, that’s where their heads start to explode. 

Debra Chantry-Taylor  19:36 

You know, it’s delicate and elevate on steroids, isn’t it? So, because if you actually get this right, it’s not about you losing your job, it’s about actually freeing you up to do the things that add a huge amount more value to the business and are probably more stimulating for you as well. Because I’m guessing a lot of the stuff that AI can do is the boring, repetitive stuff that none of us really wants or needs to do. 

John Munsell  19:58 

Well, I mean, there are a couple of examples. As I can give you. So there’s a book called The corporate life cycles by a guy named each hack a diocese. I want to say it was written, I don’t know, 15 plus years ago, but the principles hold true. It’s one of the more complicated books you’ll ever read. But man, I’ll tell you what it became my Bible as I grew our first business. And what it does is it breaks people down into four categories, producers, which are the people that just want to do the work, you know, slide the food through the door and leave me alone. I’m going to crank out the work right. Then there are the administrators. Those are the people that like the rules. They like to create the rules. They like to work within the rules. And then there are the entrepreneurs. Those are the idea guys, the innovators, the ones that are willing to take a chance and take a risk, and, you know, make things happen. And then you have the integrators, and those are the people that like to create culture, that like to get everybody together, that make a team work more effectively. And we all have some of those traits to varying degrees, right? I happen to be a high E, so I’m a big idea guy, big entrepreneur, you know, I’ve got an assistant that’s a high P A producer, which is what I need to compliment me, right? My operations person is a high administrator and a high integrator, and so he likes to build culture, but build rules around it. Likes people to work within that. You know, there’s, there’s a balance as you, as you move into the growth phase of your company, where you want, you want more producers, less administrators, more entrepreneurs and less integrators. But as you move up to what they call prime, then you want a good balance of producers, administrators, entrepreneurs and a few integrators. When you get that out of balance, when you have way too many administrators, then you chase off the entrepreneurs, and you chase off the integrators. You’re now going down the slope towards crashing and burning, right and so, so there’s a delicate balance of keeping that in play. And the reason I bring that up is we test an organisation for that. So we test all the individuals in the organisation to see what the balance is. Because if you think about it at its core, what is AI? AI is a producer and an administrator, right? AI produces, and if you give it the rules. It will maintain structure. You know, if you don’t do it right, it will hallucinate. That’s a whole different topic, right? But, but that takes that, you know, it takes the administrative roles, and it takes the producer roles and it and it makes them really fast. Now, if I am a producer and AI just took everything that really you know is, is my job justification? What do I do with that free time? Well, what we try to teach people to do is think differently. So we try to incorporate critical thinking skills so we can move them to where they they can explore new aptitudes. Or we say, look, because you’re so good at producing, you have the the intellectual capacity and the domain expertise in our organisation to drive excellence from everybody else who’s trying to use AI in the production realm. Does that make sense? So that producer is so amplified by AI that that combination is extremely powerful. So the last thing we would want to do is let that person go, because they just made AI super, super powerful. So we’re actually making them 10x who they are, and we want to spread that throughout the organisation. On the other end, if I’ve got an entrepreneur, an idea person, I can use AI to take the stuff that the entrepreneur hates, which is the production and the administration, I get AI to do that. Now, my my entrepreneur, my idea guy, can be even more creative and more productive in our organisation, but you have to know how to do that in an organisation so that you can get your best people to be even better versions of themselves. And in that case, they’re not fearful of AI. They’re welcoming it because it makes them even more productive and happier, frankly, in their in their jobs. Does that make sense? Oh, it makes perfect sense. Yeah, it’s a different way of looking at it, that’s for sure. That’s lovely. It’s really good. 

Debra Chantry-Taylor  24:49 

It’s actually really interesting. On a personal level. I’ve been trying to get my EA to to embrace AI a bit more. We use it a lot for the very general stuff, but not really. Really, in terms of taking away some of the repetitive stuff. So I’m going to make her read your book and see if she can get some ideas from that good. 

John Munsell  25:07 

Yeah, do it. She certainly will. I have prompts on the back end of the book that will teach how to do certain things. You know, one of the things that you probably have noticed if you look at LinkedIn, it’s if, if you haven’t gotten to this place, you will pretty soon where you can spot an AI post a mile away because it follows a pattern. In fact, pattern is one of the words that it really likes. So it follows a pattern. And then you’re going to see replies to your LinkedIn post that follow a pattern, and there i ai generated replies. And what what’s frustrating is when I see an AI generator reply and to an AI generator post, then what I’m seeing is AI talking to AI. And what benefit is that, other than some sort of an algorithm trick to LinkedIn, but you’re losing the human connection. And so I get more nauseated by the person who’s replying to my post in AI. I don’t really want to be connected to them. Does that make sense? 

Debra Chantry-Taylor  26:12 

Oh yeah, absolutely. I say all the time, and I often block people because of that. But can you My question for you? Can you train AI out of looking like AI, because you said you can tap into your expertise and your experience, and you can teach it what you do. So is there a way to actually break that pattern and create something that is tapping into your because I’ve used it for writing more around sometimes generating ideas, usually kind of taking my very verbose stuff and breaking, bringing it down to a lot, sort of more concise, but also to challenge certain things as well. So I mean, you can use it in many, many different ways, but can you break the pattern that it has that people won’t know it’s AI? 

John Munsell  26:55 

Yes, you can. So there are a couple of ways to do that. First of all, what we do, you know, we have what we call containers, and so I’m going to give a container around the style of voice that I want inside of that container are variables. So instead of saying I want you to talk in a casual voice with a little bit of humour and blah, blah. I’m actually using variables to tell it specifically how to calibrate it. So I’m using a variable about sentence length. I’m using a variable around paragraph length. I’m using paragraph length, I’m using one around humour. So humour equals three out of 10, right? Three, slash 10. All of the AI tools know what that means. Three out of 10. They know what that means. So if the thing gets off the rails and starts being silly, I’m going to dial down the humour level, and I’m going to crank up the tone level. So we have a tone 77 out of 10 those things. So we have these different variables. I talk about that in my book, but that is like the difference between the old stereo used to have in your car, where you just had base in trouble, to now having a mixer board where you can start to dial things up and down so it sounds a lot better to you. The other is providing it with your own thoughts, as opposed to just telling you to go write me a blog post about X, you teach the AI in your prompt to have a back and forth with you, to push you to challenge your your own thoughts. And then you start to integrate your thoughts into it. So you’re now, instead of telling it to go write something you’re having a dialogue with the AI back and forth, and it pushes you to think a little bit differently, and you can push back, and then it’ll compose something that sounds really good. One of the other blocks in our AI strategy canvas, the eighth block. There are nine blocks in it. The eighth one is the rules block. So we teach people how to create rules that tell it you know what not to do, right? Don’t use certain words. They’re getting better at what you would call negative prompting, but you can create a list of exceptions. Don’t use this, right? Don’t use the word Delve. How that used to be a big word, right? They would always use they’re getting better at that. But what’s cool is, the better you get at these, at identifying patterns, the better you are at changing your rules to break AI out of the patterns. But you will only see it when you’re starting to write stuff and you’re like, oh, this sounds great. And then, but you have to, and this is where discernment comes in. You have to be able to read what it just wrote, and go, Okay, this sentence is 15 words long. It doesn’t need to be. It sounds good. In other words, it sounds intelligent, but when you really break it down, it it’s not making sense with the rest of the context. In the sentence, you know, it uses a phrase that hasn’t really identified what it’s talking about very clearly. And you can kind of see that pattern as you go through a lot of writing, you can see that. But all of those things just mean the AI is missing context and direction. And when you learn to feed it with the proper context and direction, what it writes, man, it’s it’s good, but, but not all tools are are for the same thing, right? I love Claude. Sonnet 4.6 for writing, does a great job. I am enamoured with claude’s Opus 4.6 I think it’s amazing. If I’m writing business stuff, I’m going to go to chat GPT 5.2 I’m also using a lot of Gemini, depending upon what I’m doing, if I need real time information, I’m going to grok and perplexity, you know. And what I typically do is I realise how fast it will write, but I use multiple llms to produce all this. So I’ll do research in perplexity and pull back a lot of information. I’ll do some research in grok to pull back more information. I’ll take those artefacts and then I’ll put them into a project that I have in Claude That massages all that and then asks me for my thoughts. What do I want to put into that? What are you know? Do I agree or disagree? Do I want to challenge what’s in that article? Or how do I want to synthesise it? When you do that, you know you’re, you’re, you might take a little bit more time, but what it produces is so more so much more meaningful and thoughtful. You know, you’re not regurgitating what everybody else is. You’re now integrating information, and you’re informing people. And so if you want to know how to get AI to recommend your site, you know you now, you know, everybody’s talking about getting ranked in Google search engine. Now, it’s really cool if, when somebody does some research inside of AI that they point to you as a resource, the way to do that is to write meaningful content that acts as a resource, not as a regurgitation of the obvious and that, and that’s how you do it. Does that make sense? You do your research, first from other tools, pull it together, and then you compose what you’re trying to compose. 

Debra Chantry-Taylor  32:52 

I must have done quite a fair bit of work on this, on my own things. If you put in best EOS implementer in Australia, you’ll very quickly find me there, which is good, yeah, but it is interesting. I don’t know, do you find yourself fighting with your llms at times? I mean, I I get into some really heated debates, and I’ve threatened to send her to Siberia at times, because, you know, you ask her not to do something, and it’s like she just keeps doing it. It’s like, I’ve told you, do this one more time. I’m sending you to Siberia, and we have a bit of a joke now. But, yeah, yeah, 

John Munsell  33:21 

I used to write I like, for instance, the M dash. I like, Would you quit with the M dashes? You know? But I like me a good M dash every now and then. 

Debra Chantry-Taylor  33:30 

The M dash is a beautiful piece of grammar, and using the right context, it was just overuse. That was a challenge, yeah, 

John Munsell  33:37 

but not every other sentence, right? And that’s, that’s the frustrating thing, but, but, yeah, I used to get frustrated with it, until I learned how to create the right, what we would call a prompt stack. And I create that prompt stack, and then it becomes the system instructions inside of a Claude project or a custom GPT or Gemini gem, or something like that. So all of my writing framework is inside of the project, and all of the knowledge based documents are in that project. I probably have 100 different projects inside of Claude. I probably have 20 Gemini gems. I probably got 150 or so inside of chat GPT. And they all do different things, but they, the struggle, though, is they all they have overlapping knowledge based documents, and so if I update one document, I now have to remember, okay, I got to go change it in 75 different places, and that’s, that’s frustrating, that must get better over time, surely. Oh yeah, yeah. Well, they’re gonna, they’re gonna figure this out. I mean, like, I know that if, if you’re a Microsoft 365 user, you can integrate with OneDrive and. A SharePoint. And so then you could use copilot as your mechanism, and it will you change that file once inside of your SharePoint folders. And you don’t have to worry about all the different places that you’ve you’ve stored it, but for the rest of us who are using the other llms, you got to track it. But if you’ve ever, have you used claude’s co work? No, I haven’t. No, okay, are you on a Mac or PC? I’m on a Mac. Oh, fabulous. Then you can get it. So co work is, is fantastic. Now it works on your drive, okay, so it’s, it’s an app that you can download, but it literally works on your drive, and you can tell you can tell it which folders to work within, but it’ll organise things like, I what, here’s what I did. Because, again, I had, like, 75 different Claude projects. I had to figure out where all the files were. So I fire up CO work, and I said, I want you to go into all of my Claude projects, and I want you to make a list of all of the knowledge based documents that are in them. And I want them to have a date that shows when they were created, because I’m very careful to make sure they have version dates in there. And then, you know, through a couple of steps, I said, I want you to tell me which, which of the files in which of the projects need to be updated, because it’s going okay. There are eight versions of this document. This one appears to be the current one, blah, blah, blah, blah, and it would do it. And so now the next step was, okay, the source documents are in this folder, this folder, this folder and this folder, I need you to organise all of those archive, the ones that are out of date, create a single folder with all the ones that are the most current, and then tell me which you know things I need to update it when I’m and I’m not exaggerating, in 20 seconds, it built all the folder structure, organised everything, moved all the old ones into an archive folder, all on my hard drive and like, that’s insane. And so that, then it gave me the directions to tell my assistant, okay, all of these projects need to be updated and made current, and we need to change the naming convention of the files, you know, blah, blah, blah. And actually, I did that inside of CO work too. Now that I think about it, I changed the naming conventions of the file so that the date was up front and blah, blah, yeah. Anyway, it was amazing. 

Debra Chantry-Taylor  37:43 

I’m excited. There’s so many things I can see for potential of it. I am curious, though you mentioned the three out of 10 humour. I’m wondering, if you go 10 out of 10 humour, does it become dad jokey, or how does it behave? 

John Munsell  37:55 

Oh, yeah, yeah. So okay, then you get into something called prompt conflict. But, but let me. Let me give you an example. One of our students was building a chat bot for ophthalmology clinics. And so the idea was it would train the employees of the ophthalmology clinic How to Sell ophthalmology services, right? So they would interact with this thing going back and forth. She comes to, we have these things called office hours, so our training is recorded, but every three days a week, we have Live Office Hours, and you can come bring your problems to us. And so she said, John, I need your help with my chat bot. I’m like, Sure, what’s up? She says, I can’t get it to stop talking like a pirate. And like, and I’m like, a legit pirate, right? And she goes, yeah. And I said, Well, show me, because I was like, maybe she’s, you know, making it exaggerating. And so she prompts it, and sure enough, it goes, arm 80, you know? And like, oh my gosh, this is incredible. So I said, All right, let me look at your prompt. Well, she didn’t use the container method with the variables. She just used a container and says, I want you to use a little bit of humour. And then she said, Oh, at the beginning she goes, You’re a helpful assistant that’s teaching our people how to do X, Y and Z. And then down below it, she says, I want you to use a little bit of humour. Don’t be so dry, be be more humorous, kind of like that crazy uncle. Okay, that’s nutso. So think about what she just said. Okay, she said, I want you to use a little bit of humour. Because originally she she said, I want you to use some humour. And then she said, a little bit of humour. And then she would say, she would say less. She couldn’t get it to calibrate. But I said, Look, your problem is that you said kind of like that crazy uncle. The uncle part tells the AI, it’s talking to a child. For some reason, it thinks that children. Thinks pirates are funny, but because you said it’s a crazy uncle, it’s going to go into this crazy ring and and it collides with the helpful assistant part as well. So she had a chocolate mess going on. So we went back to the drawing board, and we said, Okay, we’re going to knock down the humour to two out of 10. We’re going to give it a more directive about what the tone is, what the style is, and all that stuff, using the variables. Cleaned it up in literally a minute and a half, and it was talking normal, but she didn’t, who would have thunk right? 

Debra Chantry-Taylor  40:39 

Yeah, absolutely. Yeah, I can imagine there’s some real horror stories out there around people getting these things wrong here, sure, 

John Munsell  40:46 

but you know, if you, if you did crank it up, just to your earlier question, if you did crank it up for 10, but if you told it, I’m writing, I’m writing a medical brief for a medical journal. I want humour to be 10 out of 10. It would, it would see that as major conflict. Whoa. No, this medical journal, and you’re, you’re a brain surgeon, or whatever, it’s going to go, okay, 10 out of 10 for that, that’s where the conflict comes in. It would then make it more familiar, but it might make an occasional quip, but it wouldn’t go crazy because of that conflict. On the other end, if you said, I’m going to write an email to a friend and I want you know humour to be 10 out of 10, yeah, I’d go it got bonkers on you. Every line would try to be a one liner. You could also throw in their sarcasm. Equals nine out of 10 and watch what happens. It’s really funny. 

Debra Chantry-Taylor  41:48 

I am going to have some fun with this now, for sure. Okay, we could probably talk about this for hours and hours. But what I would love to know is what has been the sort of one of the biggest efficiency gains you’ve seen in a business 

John Munsell  42:00 

in ours, or any business, any business, no, I think that you know, the biggest one that we personally have accomplished is I did something the other day that literally would have taken me three weeks to do, and that was, build an application. Okay, that paei stuff I was telling you about, I needed to test an organisation that has 12,000 employees, but I wanted a dashboard where they could visualise how their organisation was structured. I wanted them to see that their pa ei formula against the the 10 levels of AI mastery. And, you know, having been in the software business for 27 years, I had a team of developers. What I envisioned was I needed to send out a survey to everybody. I needed to really calibrate the heck out of that survey. I needed to get those results back in, and then I needed to create a report for everybody who took the assessment. I wanted them to see their own profile. I wanted AI to analyse their data and give them a nice narrative that explained where they were, but then I wanted the owner of the company to see a dashboard so they could visualise in the heat map, where their people were in the 10 levels of mastery, and where the the hot spots were in the PA ei framework. Okay, that would have just writing the requirements document for that would have taken me two weeks back and forth with with two developers building that as an application, that somebody would go take the survey. I mean, writing the survey would have taken three weeks, building the survey, letting somebody take it, giving them the display results when they’re finished, emailing them a PDF that’s nice, neatly laid out, emailing me a copy of the PDF, creating a database of all the responses, creating a dashboard with heat maps and all that stuff that would have taken six weeks. I probably would have charged $70,000 to do that. I built that entire thing with AI in three days. And it was unbelievable. I mean, it was just amazing how it works. So, yeah, you want to talk about an extreme example. 

Debra Chantry-Taylor  44:31 

That is an extreme example. Yeah, it all comes down to No. It’s like everything, isn’t it? It’s like shit in, shit out. You’ve got to make sure that you actually know how to use and how to get the best results from it. Hey, I’m saying we could probably talk about this for hours and hours, but unfortunately, we haven’t got all the time in the world. I’d love to you to give us the top three tips that you would give somebody who’s considering taking AI beyond where they’re probably at. As you said, level one, level two, maybe level three, up to a five or a. Six. What would be the top three tips you would say to them? 

John Munsell  45:02 

Well, the first thing I would say is, don’t, don’t get enamoured with the idea that you’re going to get it all in one place. In other words, you should have in your arsenal multiple llms. And I would probably not do fewer than three. So I would probably have chat GPT, I would probably have Claude. I would probably have either Gemini or perplexity. I’m probably going to lean more towards Gemini, because Gemini does images, does videos, that that whole Gemini suite is amazing, so I would say, you know, make sure that you have paid versions, not free versions, but paid versions. So those, those three combined, will be 60 bucks. What does that give you back in terms of time? You know, if you can’t, if you can’t recover $60 then there’s something wrong with you. Mean, but yes, in your job, that will that combo will save you probably eight to 10 hours a week, but so so I would, I would make sure that you get multiple models. Then I would say, you, you you need to quickly move to where you’re not just having a question and answer back, but you learn how to structure prompts so that when you go and see them again two months from now, you know what it’s doing, right? And so you want to learn how to structure the prompts in a methodical fashion. You want to learn how to save the prompts so that you can use them later. And then you want to get to the point where, once you produce an artefact, like a report, like a research report, or something like that, that you’re storing that. And then you now move up to another level where that becomes a knowledge based document in a project, or a GPT or something like that, so that your project has system instructions that are based on that the what we call scalable prompting. So, you know, you can calibrate it, and you can organise it, and then it has all that information in there, like, you know, when we’re developing your proposal, our pricing is in the in there, our description of our persona and our company and all that stuff is already in there, so I don’t have to do it over and over again. So that would be the next thing you know. Quickly learn how to build these tools for yourself, but make sure you understand how to build a knowledge based document as well as a prompt that can sit there as a set of system instructions. And then the third thing is, I would make sure that you learn how to to get AI to uncover things that you would not normally uncover. Like, for instance, in our meetings, we all hold, we hold hold, 90 plus percent of our meetings over zoom. And if we’re not holding them over zoom, and I wish I had it here, we’re holding them live. And we use, what’s it called, a plot note. I don’t know whether you’re familiar with applaud note, P, L, A, u, d, it’s, it’s the size of a credit card, but it’s twice as thick. It’s it’s an AI recorder, and so we’ll take it into a meeting and say, Do you mind if I record the meeting? They’re like, No. And so we just turn on the recorder. It provides transcript, just like our AI note takers in zoom do. So we have a transcript. All of these tools give you a default way to analyse the transcript, but those are optimised around creating a to do list and a follow up. You want to customise your own prompt so that it it. You add some things to it, and one of the things that we always add to ours is, I want you to identify the things that would have escaped my attention that will add value or lead to something else in this conversation. And so you’re you’re asking it to look for the nuance that would that normally would be overlooked, and what it comes up with will blow your mind. Debra, I mean, it’s amazing, like if you have an hour and a half conversation with somebody, and you get it to look for the nuances that would have escaped your attention, it’ll sometimes come up with and say, you missed this is an actual partnership opportunity, not a sales opportunity. I’m like, Well, really, it’ll tell you all that stuff, or it’ll also tell you this person is really highly focused on numbers. So make sure anything you present covers those numbers. You know, I could go on and on. There’s a million different things that 

Debra Chantry-Taylor  49:52 

you could use, but I know I can imagine. So, I mean, there’ll be a transcript, obviously, of this, this podcast, and I guess people can just copy and paste that and use that in. Their in after their meetings. I have heard of the little note takers. I mean, they’re kind of a modern day what do we used to call the dictaphone, isn’t it? But it’s linked it directly into AI, which is fantastic as applaud note. Okay, so got to have at least three llms, all paid versions, not negotiable. And like you said, if you cannot save enough time to pay for that 60, $70 and you’ve got doing something wrong. You’ve got to learn how to structure your prompts, save your prompts and create knowledge, knowledge based documents that can actually be re referred to so you’re not reinventing the wheel every time. And then use your AI to learn, learn about yourself and your the way that you engage, interact in your meetings. 

John Munsell  50:39 

Yeah, uncover stuff that you would have overlooked Absolutely. 

Debra Chantry-Taylor  50:42 

Yeah, gosh, there’s so much there. I mean, I’m sure I might actually invite you back against I’m sure there’s a whole lot more I’d love to talk about. But if people do want to find out more, obviously, you’ve got your book, which is the engrave book, and you said there was a course as well as that, right? Yeah. 

John Munsell  50:56 

So they could go to ingrain.ai or they could go to bizzuka.com, B, I, z, z, U, K, a.com, so we have our courses there. We do a lot of corporate training. So you can connect with us there if you want to go get into some corporate training. And you know, like I mentioned earlier, we have implementers all over the globe now, so if we can’t service you directly, we can put you in touch with someone who can. So yeah, that that would be it. I would say, go to bizzuka.com first. The other thing I would do, so I’ll give you this one other thing, Debra our My goal is to write the second version of the book. So it’s not a new version of the book, per se. It’s where, after working with companies for the last year and a half, where people need to go, after they implement the things in this book, there’s a whole new thing that they need to be working on. I’m in the middle of writing that book, it should be out, hopefully by the end of April. Famous last words, though, last a whole lot longer than it did, if they will go to ingrain.ai/ 2026, which is my my fuse that I’m putting on myself to make sure I get that book out. They can register there to get on the waiting list, and if they tell me that they heard about this through your podcast, I will give them a free digital copy of the book when it comes out. So that would be the last way they could get in touch with me. 

Debra Chantry-Taylor  52:35 

Beautiful, and that sounds great. Thank you so much for that, and obviously the implementers that you’ve got around the globe there. So that’s for people like me who love to outsource everything, isn’t it? So if I don’t want to learn all this myself, I can actually have somebody come in and actually help me to get Jen up to speed, to get the team up to speed, to sort of help us put this stuff into place, to make the most of it. That’s correct. Yes, I’m liking that idea very much. Hey, John, lovely to see you again. Thank you so much for your time and for sharing all your knowledge. I’m sure I will speak to you again. I’m looking forward to sort of, yeah, hearing it, because it’s just going to get better and better and better as time goes on, isn’t it? Oh yeah. 

John Munsell  53:13 

It’s like we said, today is the worst AI you’ll ever see. Well, I It’s been a pleasure being on your your show. I appreciate it the great seeing you and connecting again. 

Debra Chantry-Taylor  53:26 

Thank you very much. It’s been a pleasure having you too. You. 

=======================================================================================================================================================

Debra Chantry-Taylor 

Certified EOS Implementer | Entrepreneurial Leadership & Business Coach | Business Owner

#betterbusinessbetterlife #entrepreneur #leadership #eosimplementer #professionaleosimplementer #entrepreneurialbusinesscoach

Certified EOS Implementer New Zealand

Certified EOS Implementer  Australia

Certified EOS Implementer UK

Certified EOS Implementer NZ