← Mondays with Matt
Mondays with Matt
AI for Business Productivity: Shaping the Future of Work
00:55:40
How AI is Transforming Work and Productivity AI is reshaping the business landscape, helping organizations work faster, smarter, and more efficiently. In this episode, we explore how AI-powered tools can streamline workflows, automate repetitive tasks, and free up valuable time for strategic growth. Whether you’re an entrepreneur, executive, or team leader, understanding how to implement AI in your business is critical for staying ahead in today’s competitive environment.
Loading summary
Transcript142 lines
- [00:02]
A
You're listening to the Build you'd Business podcast, powered by Turnkey Coach, where we help business owners find freedom over fear.
- [00:09]
B
I'm Matt Reynolds.
- [00:10]
A
And I'm his brother, Chris Reynolds.
- [00:12]
B
Join us as we help build your business and move from fear to freedom together. You're listening to the Build you'd Business podcast with your hosts, Matt and Chris Reynolds. I'm Matt, he's Chris.
- [00:34]
A
What's up?
- [00:34]
B
Welcome to the show. Glad you're here.
- [00:36]
A
Glad to be here.
- [00:37]
B
Been a busy couple weeks, I think, for both of us. And so we thought it'd be a great time to talk about one of the things that have been a huge time saver for us and that is, I think, still an enigma for many founders, and that is how to leverage AI in their businesses to save them time, to get much more done in less time. And really how you do this. And so you're the tech guy. I'm using it every single day. My tech team is using it a lot more even than I am. I certainly am using it a ton for SOPs and things like that. Still kind of in the. In the LLM model. But how are you using AI in the business?
- [01:15]
A
Yeah, so it's been a crazy shift in the last and I think it's going to keep happening. But over the last couple months, I've noticed a substantial shift where there's the more of these tools that you've been working with, the more that you're aware of what all is possible, the more you're capable of doing. And right now, it's still so early for so many businesses that if you're on the early side of this, you're like way early. Like a lot of people are. Still have no idea. Most people have heard of ChatGPT, sure, maybe they've used it a little bit. But in terms of the all the tools in the ecosystem that are possible to use, you know, PE people just have no idea. And so I just figured that we would take today to talk a little bit about how I'm seeing this used effectively by different founders. Today, for example, I will literally have meetings all day today. 100% of them are about me showing or teaching people how to utilize these tools in an effective way. I am at this point, it's all I do.
- [02:21]
B
I also have meetings back to. I had eight hours of meetings yesterday, nonstop. Eight hours of meetings today, nonstop. All of them were set up by AI. I set up nothing. And everything goes into our CRM in HubSpot. I can pull up their contact Information, I can see things about them. So I'm going into a call, knowing who I'm talking to, knowing some of their background and we're using it every single day too. Certainly not as much as you guys are. So the one thing I would say you were talking about that everybody is familiar with AI at this point. I know that AI is also one of these things that are kind of scary to some people, like, where's it going to go AGI? Like, are we going to have robots all in our house in five years? I don't know. The key here right now is what it's able to do today and how to continue to leverage that. I think, yeah, certainly there are people, those who are running the major AI companies who are. They're not thinking about today. They were thinking about today 10 years ago. Yeah, they're thinking about 10 years from now and 20 years from now and. And you're probably thinking more forward thinking than I am in this. I'm just like, what tools are out there to make me more productive today? And so, so I would just say that if you're not using it yet, it's not hard to use, but it's just like anything else. It just takes some time to play with and you'll find that you'll get way, way, way better really quickly. This is the other thing. It's sort of like when we were kids and you get a new Legend of Zelda game and you're terrible at it in the beginning and you play it a bunch and within eight hours, 12 hours, whatever, you're like, oh, I'm actually really pretty good at this. Now that's this thing. So I would encourage any founder, even if they're listening to this and they're not running a tech company. And I'm now in this weird space where we started as a service business. We now run a SaaS business with the B2C service side. We're using it in both. We're using it a ton on the SaaS side, but we're also using on the service side, the people management side, the org structure side, the SOP side, all of those things. And that's surface level stuff. But that's a great place to start for founders who haven't used it so far because you'll start to find like, whoa, this thing is like amazing. I can do all kinds of cool stuff on it.
- [04:38]
A
Absolutely. And we'll talk a lot about how there's a process you can kind of follow to get familiar with AI, start using different Tools start getting better at. And you're exactly right. It's like anything else, you have to play with it to get good at it. So everything that you do works that way. So this isn't really different. It's just in some cases I think it's a little scary. In some cases, people don't know how they're going to use it when they start. So they don't know what to do and they're like, ah, I just won't start. Well, that's a bad idea. You want to start. So, so let's just do a little bit of background here. If you're a longtime listener to the show, you know at least a little bit about me, but I'll give you a little bit more background than maybe what we've done before, at least on the tech side. So I've been doing this for about just over 25 years in terms of software engineering. Deep background in all the areas of technology that you can kind of shake a stick at. I'm super, super nerdy. I'm way nerdier than most people think I am. I love everything about the tech stuff. I love to learn about it. I like to play with it is my natural inclination to do this. So when AI stuff started sort of emerging, we saw the early versions of even the paid accounts for ChatGPT, for example. I was probably buyer number seven or something.
- [05:51]
B
We were super early too. For several years we've had it paying the 20, 25 bucks, whatever it was on the very first one, and started playing around and testing it out.
- [06:01]
A
So when they came out, you know, there's no AI tool that I'm aware of that is at least in that major tier of AI tool that I don't have access to. I have and I play with all of them. I do various things in them. And because I do this so much, I think I've got decent insight into places where even non technical people can use these tools in a way to just massively accelerate your productivity. So we're going to just talk about what this is and sort of how do you get into it? Maybe we'll start with this idea of we're not going to talk about AI at like a scientific level today. We're not going to talk about anything that's too nerdy. We're not going to get anything of that. The nerdiest we're going to get is there's a particular tool that I think might be interesting for those of you that are a little further on the tech. The Tech end of the spectrum. We'll talk about that towards the end. But as you think about the various ways that you can use AI, I think it's, I think what is missed by most founders. It's crazy how frequently I see this. Matt, I'm curious if you see this too, but you basically get people that think they're supposed to integrate AI into their product as like a first step. That isn't it at all. That's not even close to right. The thing that you're supposed to do with AI right now is you can use it for all of the internal things that are annoying and slow and things that eat a lot of time but don't drive a lot of value in and of themselves. Maybe they drive value in the long run, but they're not something that. I'll give you an example. One of my favorite examples, and everybody started using AI for this early on. It's wordsmithing. I don't know. That's my phrase for we're trying to figure out how to get the words exactly right in terms of marketing or writing blog posts or emails or whatever it is. Right. If you look at a pre AI era, how much time most people were spending on wordsmithing that care a lot about the context that they're writing in now? A lot of people, I get a lot of emails from people who do not wordsmith and kind of wish they would wordsmith a little more.
- [08:03]
B
Yeah, right.
- [08:03]
A
But. But in this particular case, you know, you're. You're talking about a ton of time spent on something that while it does put polish, it doesn't drive a ton of value. Right. And so you want to be able to do this in a way that is effective. You also want to be able to do it in a way that is personalized. And so we'll talk about what does that look like and how do you make it do that kind of thing. But that's the world of opportunity that is available, and it's much bigger than you think. Big, because many of the skill sets that have been required up to this point, that would have taken an enormous amount of time to learn, are now available to basically everyone. And most people just don't realize it's there. So, you know, there are people who find out that there is a tool out there that will allow them to design very quickly a proof of concept application, and it looks super good and they can deploy it in a way that, you know, it's not production ready, but it's the kind of thing that they could take out. And if they wanted to show investors, if they wanted to show potential clients, this is kind of my idea. This is what I'm thinking about doing.
- [09:11]
B
It's like pre mvp, but you can at least see what we're talking about. Yeah.
- [09:15]
A
And in some cases, they're beautiful. I mean, they look really good. They're not ready to take on customer data and stuff like that yet. Someday, I suspect they probably will. But until then, what's happening is the gap between your idea, your thing, and your ability to get a ways down.
- [09:33]
B
The road on that thing, to communicate that thing effectively.
- [09:37]
A
Yes.
- [09:37]
B
Is shrinking. And to come back to the point that you made, I think one of the best places to do this in the beginning. The best place to do this in the beginning is internally.
- [09:45]
A
Yeah.
- [09:46]
B
So say I've written a bunch of SOPs over the last few days. In the past, I would write a sop. I'd spend a lot of time writing the sop, get the draft. But the first draft was pretty good. But then I would spend a lot of time working through the draft and doing the second draft and the third and getting the verbiage exactly right in the copy. Now I can write a quick and dirty draft as quick and fast as I can run it through ChatGPT and say. And as long as I know how to. Or grok or. Those are kind of the two primary ones that I'm using right now. And have them take that draft and say, prompt draft. This is what I'm trying to communicate to my staff. I want to do it in a way that is tight, clean, clear, memorable, et cetera. Go. And then they do it and they spit out something that is way better than what I wrote. And then it usually doesn't stop there. Then I say, that's great. Now, what if we took what you just wrote and then we actually changed this about it? Go. And then it goes, and it's. In one second you have it again, and the next thing I know, I had. So we were just at the Arnold Classic, for those of you that are in the fitness industry. Had nine people at Arnold Classic. Three or four of us were salespeople. But for all intents and purposes, I mean, it was very expensive to be there. We had a super nice booth. Our booth looked like a IKEA living room. Right. And in the world of fitness, everything is like black and red and blood red and heavy metal, and we just stood out.
- [11:14]
A
You're like, we're going soft.
- [11:16]
B
Yeah. It was perfect because we were trying to draw in Professionals, not, you know, uber tanned bodybuilders with string tank tops on. That's not. We specifically didn't want them coming to the booth. Okay, so the capturing the view was there. The question is, how do I train up five more people to be able to do a, what we call a disco call or a discovery call, ask questions, hear the pain points, hear the problems, solve the problems, walk through. A demo demo of the software is not a tour of your software, especially if it's a complicated software, because it just blows everybody's mind. Your goal is to just solve the problem that they identify as the pain points. But I needed to do that in a way that trained the staff up very, very quickly in a way that was memorable for them so that they didn't have to go back and go like, wait, what was the next step? This is where AI was so perfect because by the end I had a half sheet, literally a half sheet of paper. Like that's the process. And then we would role play it back at the Airbnb very quickly. They could remember the five steps and go and go. Right. And so that's how we, and that's a, that's. I would say one of the simplest way to, ways to use AI is in that wordsmithing piece like that. It actually created an acronym for us. We use C, L O S E. And we had, you know, words attached to each one. That was perfect because it was easy to remember. So it wordsmithed it perfectly for us. So we were, we were doing this all the time. Now post Arnold, I've calls all day about how we do these follow up calls, demo calls, how to close, same thing, right? So this is the goal. These are the bullet point action steps, kind of quick and dirty draft. Now clean it up, make it clear so that I can present it to the staff. And then you just tighten a little bit and it literally cuts the time of writing those very clear, very systematic standard, you know, operating procedure pro, like process and procedure. It cuts it in by about 90%. Yeah, it's crazy.
- [13:18]
A
Absolutely. It takes one of the most critical parts. I mean, you know, we've had a lot of podcasts around the idea of building standard operating procedures and the value of building standard operating procedures. Frequently that is what gets missed, right? Everybody's too busy, too busy to do it. You're no longer too busy to do it. Because if you simply gain this skill and yes, it's going to take a little time. Okay. It isn't something that you're going to be perfect at when you first start. In fact, you're going to start probably a little frustrated. Why does it not know the thing? And you'll learn very quickly. Something that I'm watching people learn as I'm training them.
- [13:51]
B
Yep.
- [13:52]
A
I'm watching them learn that they are not very clear in the thing that they say.
- [13:57]
B
Right.
- [13:57]
A
So it's the, like, that's the thing. Right.
- [13:59]
B
And then how to prompt the AI to then make it clear. Right. With the right prompts. And interestingly enough, and I think this is becoming more and more available to all, but for somebody like me who's put out thousands of podcasts and written documents, it's. I have stuff all over the Internet. I can now say, now take that and put it in Matt Reynolds, Personality and voice.
- [14:23]
A
Yeah.
- [14:23]
B
And it knows my personality and voice. And it uses terms that I've used for years that I forgot I would not have put into the sop. And it spits it out and it doesn't sound. Those early days, the early permutations of AI, you could look at it and say, this is AI.
- [14:39]
A
Yeah.
- [14:40]
B
Now we're at least in the LLM version of this, we're now passing the Turing Test where it's like, it looks like they really wrote this now.
- [14:49]
A
Oh, yeah.
- [14:50]
B
Let me be also be clear. I wrote every word. Me and my editor in chief of the book of Undoing Urgency. We did not, we did not use AI at all for it. It needed to be my story, very personal. So that took some time. If I go back and do it again, maybe I still would run it through and just get some input on, you know, I'm not crazy about these three paragraphs.
- [15:09]
A
Let's.
- [15:09]
B
You could do that as well. So.
- [15:10]
A
Well, look, even the workshopping of the outline. Right. Shopping and outline, it's great to use that. And you can do all kinds of things. We'll get into some of these on this one today. But you can do all kinds of things to bring into the equation your. If you, you know, most of you probably don't have a ton of published stuff out on the Internet, the LLM is likely not trained on your voice. And so there are very simple ways now to provide it with your own stuff so that it does sound exactly like you.
- [15:42]
B
Yeah.
- [15:43]
A
And we'll kind of go through what some of that stuff looks like. But these are the tools.
- [15:46]
B
Right.
- [15:47]
A
So the first thing that I have to explain to people is, you know, if you're a founder or even if you're a top level exec at a, at a Startup of some kind. And you're thinking, how am I going to. How should I be using AI in the product? You're not thinking right about it yet.
- [16:02]
B
Right.
- [16:02]
A
There is an ab. There's absolutely valuable stuff in how do you implement AI within your products? That's true, but no one's ever done that well who doesn't know how to use it for all the other things.
- [16:14]
B
Correct.
- [16:15]
A
And so you start at a place where you're literally just using it for this very basic, normal, internal stuff that you have to do. It's the drafting of emails, it's the laying out of I don't know how many of you ever get a list of things that you're supposed to get done. And in that list you become a bit overwhelmed with everything on the list. One very interesting trick is to provide that list to the right LLM and say, I'm super overwhelmed and don't know which order to do this stuff in. Can you please help me figure out what order to do it in? It'll come back and tell you what order to do it in. And it will be pretty close to right right now. It'll be better and better and better depending on how much context we give it. And that's the key with AI. So as we think about how to make AI more and more and more effective, I want to talk about it in a way that I think everybody can kind of visualize, primarily because we are doing this on a podcast and most of you are listening to this, not watching us yet. This is on YouTube, by the way. You can watch it on YouTube.
- [17:10]
B
You should be watching it. It's better on video.
- [17:12]
A
Well, you should be watching it unless you're driving, in which case you should not be watching it. But good call. But in general, yes. So when we think about AI, I want you to think about that ChatGPT interface. If you haven't seen it before, I'll describe it. The basic idea of Chat GPT is you have this big open text box. You can type in it. You're typing to the AI like you're typing to another human, basically.
- [17:35]
B
And you know, it's very reminiscent of old school chat. Like, you know, like AOL chat, whatever.
- [17:43]
A
Like the Instant Messenger. Instant Messenger. That's what it was. Yeah.
- [17:46]
B
So it's a messenger and you're just. But rather than a human on the other side, it's a computer on the other side, but that has all the information in the entire world to spit back answers to you.
- [17:56]
A
Right. And you want to like, I feel like you. One of the perspectives on this that I feel like is helpful is to remember that every major AI today has been trained on every written word humans have ever made in all of written history that we have. You know, obviously the Library of Alexandria we don't have. So not, not everything is in there, but everything that is recorded, it's been trained on. And I always find it interesting when, and this is probably my tech background coming out where people will say, ah, this AI is so dumb. And I want to be like, actually what's going on here is that you're prompting it very poorly. And it's happening for two reasons. One of them is that you're not communicating very effectively to it. And the second one is that you have not given it nearly enough context to know the thing that's in your head. Believe it or not, those two things are also the reason that people are ineffective with their employees. It's the same, same thing. So that AI is behaving in many ways like a, a low level early employee. And if it doesn't understand what you're saying, neither do your employees. So the first thing you gotta learn is how much context do I need to give this thing in order for it to understand what I needed to understand? And the answer is quite a bit. So as much of it as you can possibly get is going to be really, really valuable, what that looks like. And one of the ways this can be really, really effective for early founders especially is, you know, most people. Now, I think we've talked about the, the tools that you use on your meetings. If you have a lot of meetings, you're probably using an AI note taker. That's right. So people are like, great, we use Fireflies.
- [19:31]
B
So Fireflies fl. Full transcript. It will both the AI of Fireflies will give you an overview of the entire meeting so that you can read it in 60 seconds, 90 seconds, but also a full transcript. So now rather than typing the draft, you can speak the draft, put that into it transcribes, and put that into the LLM when we say ll, it's.
- [19:53]
A
Just a large language model.
- [19:55]
B
Yep. And now you have the context with which to then ask the AI to clarify, to condense to whatever it is that you want to do to be able to convey X. Yeah.
- [20:08]
A
So let's talk about what do we do with all this data that we're gathering? Because for the first time in human history, it's possible for a startup somewhere to have recorded every meeting that they've ever done correct and have tons and tons and tons of context. And there are people that would think, well, you know, I'm using AI. It's doing the. It's doing the summary. That's cool. But there's more that you can do. So I'm going to give you an example. On this particular podcast, and in all podcasts, I guess you and I take notes, right? We. We pre establish, here's what our notes are going to be, here's what we're going to do. One of the things that I did on this one, and this is getting more and more advanced in the way that we're doing it, was I got the full transcription of all of our podcasts. This is episode 16. So I had 15 episodes that I could lean on. And I have more than that if I go back to the legacy episodes, but I have all these episodes that I can lean on. I took the transcriptions and I asked AI to summarize it in a very particular way that I wanted it summarized in a way that made, you know, that made sense, had a lot of detail in it because I wanted to make sure we captured all that detail. Then I took all those summaries and I said, okay, now look over all the summaries. This has been a very popular podcast. I want you to create a new document for me, and the new document is going to be called Audience. And I want you to describe what my target audience is based on the content that we have and based on the fact that we know that it's been a popular podcast. And it was like, okay, out it comes. Here's your audience. This is what it is. I read over the audience. Spot on. It's absolutely perfect right now. I could have made it maybe slightly better if I went and found comments on YouTube, if I comments on Spotify, whatever, right? Maybe I could have done a little bit more detail. But this is all the context that we're feeding the AI in order for it to get better and better and better at the thing that we want it to do.
- [21:59]
B
So what you just described still sounds like a lot of work. How much time did that take you to transcribe, put in and get the spit out from start to finish?
- [22:09]
A
Five minutes.
- [22:10]
B
Five minutes.
- [22:11]
A
Five minutes, guys.
- [22:13]
B
You got to hear that. Five minutes to transcribe every single thing and find our target demographic. And then you took the target demographic and you did what with it?
- [22:24]
A
I did one more thing first. I said, go over our previous podcasts, the full transcriptions, and I want you to tell me every place that Matt and I Said, oh, we should have a podcast about this in the future, right? And it goes, oh, here's what they all are. But the thing is, we actually had some of those podcasts. So I said, all right, now take that list and cross reference it with the summary of all the podcasts that we've had before and eliminate all the things that we've already had a podcast on. It's like, cool. Done. That was, you know, 15 seconds, something like that. This whole process is complete. Now what I have is I have a perfect summary of all of our podcasts. I have this little index. I made it make an index where it was like, episode one, here's a description, Episode two, here's a description. And it was a two or three line description to talk generally about what the podcasts are about. And then I said, okay, now what I want you to do is I want you to look at the summary of everything that we've talked about. It's been successful for this audience. Now generate 40 new ideas for us to talk about that would be. In that. Where our, this audience would be interested in it. And it was like, there you go. Right? They're great. I looked at every one of them.
- [23:35]
B
Yeah, you, you texted it to me last night and I was like, these are. There we go.
- [23:38]
A
They're fantastic.
- [23:39]
B
There's the 40 next topics.
- [23:40]
A
Yeah, they're great. So you've got this incredible tool that's sitting there and now you can do other things with it. Now here's, here's where the real insanity starts for those of you that are already shocked, like, whoa, I hadn't even thought about that. The first thing I'm going to say is just the first little takeaway. All this data that you're generating, all your transcripts, all your emails, all your Slack messages, that stuff is all context. And if you can gather that in various ways, you have even more information that you can use to give it. Put things in your voice and to summarize things and have more awareness about the reality of what your day in and day out chaos looks like so that you can have a reasonable conversation with a disinterested third party, AKA AI, and say, I'm struggling with this particular thing, I need to figure out how to solve this particular problem. Here's an enormous amount of context about my business. Help me think about this. And you've got this little mini board of directors essentially that can go through, look at all the details. They know all the details about your business. A lot of boards of directors don't that's right. And you get a lot better information. Now again, you don't have to take the advice and it could be wrong. And if it's wrong, you come back and tell it it's wrong. No, this isn't right. Because of this. Right, right. And when you do that, have that conversation, it's really, really truthfully effective. For one thing. And here's what it is. This is the thing everybody misses. It's an incredible thinking tool. It's just the best thinking tool ever. And in combination with the fact that it's an incredible teacher, you can ask it questions about, well, why did you come to that conclusion or help provide some supporting evidence. Because I don't think that thing you're saying is true. And it'll go, oh, well, it's because of blah, blah, blah. And it provides you all this information like, oh, I just learned something today.
- [25:27]
B
Part of this, you know, in a board meeting, which we are in all the time, they last hours and hours and hours. Right. And even you're in lots of them. Ours is probably the shortest that you're in. We're pretty tight. And we're tight because of the type of software that we're using to put this stuff together. And we use AI even to build out the board decks and whatnot, send that to the board ahead of time. And they. But still you're talking about getting feedback from real humans. And with AI, you can actually get feedback not in a 3 hour, 4 hour, 8 hour board meeting, but you get it in 10 seconds, 1 second, 2 seconds. I mean, it's instant, it's feedback. And so you talk about it's a thinking tool. Same thing. I've had some great times and I would not get away from. The thing that AI cannot do is it cannot create relationship, it can't build relationship between humans. Like that's the thing that it's never going to be able to do, like again, scary movies like her and whatnot, that there will be relationships between humans and AI, but human to human, AI can't do. There have been times when we've had brainstorming sessions among my team and executives that are, they're super fun and we brainstorm, have a cocktail, get creative in the Airbnb. And we have this conversation and it's great, but what really comes out of it, we end up with one or two good ideas, but it's really a team building connection exercise. You can have the same conversation with AI and in 10 minutes come out with a hundred ideas.
- [26:59]
A
Yeah.
- [27:00]
B
And then you can even take those hundred ideas. And if you want to actually have a team building, you can take that into the brainstorming session with your team and say, here's some of the things that I'm thinking about that I've talked through with an AI, whatever, and what do we think? Yeah, okay. And then you can go into the AI and say, hey, we've discussed this as an executive team. Of the hundred ideas that you gave us, these are our top 10. I want you to reduce it down to the top three and tell me why these three are the most important.
- [27:29]
A
Yeah. And I think that this one is the strongest. Counter that point. Right. So, like, it's so powerful to have it do your straw man for you.
- [27:41]
B
Correct.
- [27:42]
A
Have it fight the other side.
- [27:43]
B
Right.
- [27:44]
A
You.
- [27:44]
B
What will the objections be?
- [27:46]
A
What will the object be? Yeah. So get this. So. So one interesting one. In the last couple of weeks, I had a. I had to navigate a tricky situation with a client. And in that particular situation, everybody runs into these. Like, these happen from time to time. But when they happen, part of what's happening is there's some communication breakdown. I think it's almost always communication breakdown.
- [28:09]
B
Sure.
- [28:10]
A
But one really useful trick here is to provide it with a lot of context about that. And as you provide it with all that context, you say, okay, what is this person looking for? What are they looking for in this conversation? What am I not understanding that I need to make sure in my next communication to them, in the next meeting I have, or the next email that I have, or whatever, what do I need to communicate to them in order for them to feel like they're taken care of and things are going well and all those things. It is so good at this. It doesn't have the emotional attachment you have to it. And so because of that, you get this incredible opportunity to use that information in. And what's really happening is, it's not like you're taking the output that it provides and you're just sending it back. What's actually happening is. And this is the part that I find so fascinating, what's actually happening is you're learning. You're reading, you're learning about you. You're learning what you couldn't see that a disinterested third party could. And in doing so, you are growing in your ability to communicate. You're getting better in your ability to manage and navigate through complex situations. That's why, beyond just being a good thinking tool, it is probably the best education tool I've ever seen. And it's to the point that for somebody like me who loves that so much that I could almost just binge on it day in and day out, it is, it is almost dangerous the amount of time that I will spend on this just because of how much, how far I can go in such a short period of time.
- [29:49]
B
Yep. Imagine the amount of time being wasted. We talk about the non urgent, non important things. People that are spending time on TikTok or Reels or X or whatever and listen, guilty is charged especially on X. But I found myself spending way more time lately on just even on my phone in my chair at night when I would be looking at X and instead I'm asking chatgpt questions or GROK questions. Right. And again, it's not perfect in the beginning. I actually like that this happened in the beginning. I don't want to get into the big political sphere of this, but there were a lot of political biases that the devs put in on this. But immediately, because there were so many millions of people using it, it was caught very quickly. And so now what you're seeing is all of these AI models are trying to move, it appears, and maybe not every single one of them away from that. So they are just factually based. And the reason that's important is because it comes back to your previous point. When I'm thinking about a thing or I'm brainstorming a thing or I'm having a meeting with employees or with staff, everyone has a lens, they're seeing that conversation through. If I can get to a conversation with AI and there isn't a lens other than the facts, then I can have a much more productive conversation, not just with the AI, but then I can actually take that and take it into the conversation with the real humans, address to real humans so it's accepted better by real humans. Right. This is because I'll say things like we joke around in our family, my staff. I don't know if your staff does the same thing. They have this term they call Reynolds exaggeration. I'll exaggerate about things. Some people would say you lie. I don't lie. Anything I exaggerate about I 100% believe occurred. Right. But it's not factual. I'm seeing it through a lens. Right?
- [31:49]
A
That's right.
- [31:49]
B
AI says that actually didn't happen, that what you think happened didn't happen. Here's actually what happened. Okay, so then how do I communicate that to this person or this staff or whatever? And so it immediately clarifies not just your it Clarifies your communication with other real humans while allowing you to learn at a rate that has been completely unknowable for all of human history.
- [32:15]
A
That's right.
- [32:15]
B
All of the knowledge of human history have been on our cell phones for, I don't know, 10 years now at this point. Right. But finding that information was very difficult. You had to really search to find the stuff. Now you can literally just go to a chatbot and ask it questions and it does the search in one second and spits it right back out to you. So what would have taken me forever to find, help me find this video. I know I saw a video one time of this, of, you know, whatever Bill Gates and Warren Buffett being, it'll find it in one second. I don't have to sit there with like, why can't I find this on YouTube? Why can't I find. And so what you're going to see, and I will see what kind of implications this makes, is that Google and YouTube, same company, have been the number one search engine for years. And now the search engine is not going to be Google and not going to be YouTube. It's going to be AI. Which is why Meta is investing so much money into AI. Why Google is investing so much money into AI. Why these companies who in the past forced you to search, you know, Google makes their multi trillion dollar company billions of dollars every single year. Now they realize not only do are we not going to force the person to search, all they have to do is ask the question.
- [33:33]
A
Yep.
- [33:33]
B
And we will give them the answer. Yep, instantly.
- [33:37]
A
That's exactly right. The takeaway from that is all the things we just said are not technical. Not one thing that we said is particularly technical.
- [33:45]
B
No.
- [33:46]
A
You can use these tools today and have enormous amounts of value back from your small investment of time and dollars. And what's more, if you don't do it, I think it's going to be difficult to start. When the tools start getting more and more and more complex. They already are correct. And there are already more capabilities than most people understand. And so it's best to start as soon as you can utilizing these tools in a way that's effective without having to worry about all the other stuff. Now, I do want to address one topic which I've had a lot of people talk to me about that are particularly concerned about, which is the question of security and information security. That comes up a lot in, specifically in business context. One of the things that I always tell people is, look, there's. When you think about AI, Most people don't realize this AI models can be run on your computer if you wanted to. Not all of them and not there are some models that are so big that you would need a giant, giant, giant computer to be able to do that. By the way, if you're an AI purist, please forgive me for some of what I'm about to say because the details, I'm not going to get into the technical details, but many, many people, I do this, I can run a lot of these LLM models on my laptop. I do it all the time.
- [35:05]
B
Yep.
- [35:05]
A
Or your phone for small, simple tasks that I just need like categorization or something.
- [35:10]
B
Yep.
- [35:10]
A
The AI doesn't, it's not inherently stealing secrets.
- [35:14]
B
Right.
- [35:14]
A
You can, you can literally, you can do this on your computer, on an airplane, disconnected from the Internet. Like you don't need the Internet if you're going to do this this way. So that's my way of saying the LLM or the AI technology itself is not evil, stealing secrets, that kind of thing. It's just, it's just a model, it's just been trained on stuff. That's it. Where you need to be careful is the services that host these things can absolutely take your data and do stuff with it. If you, you're not using a, a company that has requirements to make sure that they honor their own promises that they will not use that stuff. So I'll give you an example. Some of the largest models that everyone's using right now, obviously OpenAI is one, the company that makes ChatGPT and then a company called Anthropic that makes a tool called Claude. Those are the two most popular models. The emerging models here are Grok with X and deepseek, which is a Chinese model from Chinese researchers. We'll talk about that one in a second because I think that's one that people are afraid of and in some cases rightly so.
- [36:24]
B
Yeah. Might be one you should be nervous about. Well, yeah, but I've used it as well.
- [36:29]
A
I'll explain how to use it effectively. We can talk, talk about all that. So the idea here is, is fairly, fairly simple. There's, by the way, there's others that are lots of others that are in there and there's a ton of open source ones. But you don't need to know any of any of that stuff. The thing that matters is in the case of Claude and OpenAI, in both of those two circumstances, there is a setting whenever you purchase now, if you use the free version, they're going to take Your. They're taking your data. Right. Those things are not really free, like you're the product. So in the case where you're paying for it and you have a pro account or whatever, there's a setting that you can say, don't use any of my stuff for retraining the model or whatever. And then in a lot of the tools that use those same tools, you can do the same thing.
- [37:10]
B
So if you're trying to get financial statement output and you put in, you know, financials and you say, don't use this. Like, again, it will just pull from what it knows as a model outside world, and then we'll use the information that you give it to spit out what you need. But the information you give it, it then will not put out into the ether.
- [37:30]
A
It's not staying in there either. That's right.
- [37:33]
B
Exactly right. Whereas there's probably some nervousness around, like, Deep Seek and some of the other ones. Sure.
- [37:38]
A
And so a good example of this, when I talk about hosting, a good example of this is there's. And I'm one of these people that early on I was like, deep Seek, I'm afraid of this thing. Well, where it's hosted, it's been largely proven that the data that's being sent to the Deep Seq, you know, if you go to like, deepseek.com or whatever, whatever the website is, or use the app in those circumstances. Yeah, yeah. Your data that's going there, it is absolutely hanging out in permanent locations there in China on Chinese government servers. That's what it's doing. Like, we know it is. Like, there's demonstrable evidence. The Deep SEQ model itself, there's nothing wrong with at all. The Deep SEQ model can be deployed on a server in Amazon if you want, and you can literally disconnect it from being able to do any outbound communication. You can go inside and query it directly if you want. There's a pretty couple really good YouTube videos about a guy that does one of the Deep SEQ models on his laptop and then disconnects the Internet completely from it and goes, look, everything's working and nothing's trying to go in or out of the model. It's not. Doesn't work that way. So. So hopefully that's helpful in your understanding. The thing you care about, the thing you're looking for when you're trying to secure your communications with AI, is that the service that is providing you the AI model. OpenAI anthropic Claude Chatgpt that in those Cases that you've set a setting that says, I don't want my information used at all.
- [38:56]
B
There's a security setting there that will write. And obviously there's going to be. Some people listen to this and say, like, we can't believe any of that stuff. But also your phone's listening to you anyway with everything you say and, you know, and Siri and everything or whatever, Alexa and all this stuff, you only.
- [39:11]
A
Have so much power in that, in that scenario. And so you need to do the right thing. You need to do what you need to do and then, you know, be reasonably cautious. The last thing that I'll mention on this is just let's take my example from earlier, because I think it applies to probably every founder. Every founder is making content in one way or the other. Maybe it's email content. Maybe that's it. Maybe all you're doing is emailing customers or emailing employees or whatever. That's still content.
- [39:36]
B
Yep.
- [39:37]
A
Your content is going to be made infinitely better by the more context that it has about you. There are a million things in your brain that are nowhere else in the universe. And you have to find a way to get those things, as many of those things as possible into the context of the, of the AI tool. So when you think about this, if you are someone who's used Chat GPT for any purpose or Claude or any of those tools, and you think about this input window that you can put in, well, surely if you've done it for long enough, you've put so much text in there before that, like it says, that's too long. I can't do this. I don't know. Have you ever run into that problem, Matt?
- [40:13]
B
I. I think only one time, but, but very rare.
- [40:16]
A
Okay, so with code, which is, you know, one of the things that's been trained on tremendously, that happens a ton. Like there's lots and lots and lots of, you know, 25,000 lines limit, right? Well, well, the token limit is growing. And as the token limit grows, that's how much text you can put in there, things are going to get crazy because you'll be able to give it enormous amounts of way more context than you can even do today. But the 200,000 token window that a lot of these have today is so good that you can be really, really effective doing that. But the point that I'm trying to make here is you want to think about the tools that you use as utilizing that context window. Well, what is the context window? How much text am I putting in There. So if I have a little tiny question, what should I do in this particular situation? Right. It has no idea who you are. It doesn't know. The only way it can know that if it knows more about your values and all these other things. So you have to provide as much context as possible so that it knows how to answer. Well, when we talk about content creation specifically, and I was talking about the notes for this podcast and notes for future podcasts, one of the things that, as a programmer that drives me crazy is I don't want to go get those transcripts. We record this on Riverside and I got to go into Riverside, I got to go click a button and download it. That's for the birds. I don't want to do that.
- [41:46]
B
Sure.
- [41:47]
A
So what I'm doing now is I'm utilizing a tool that, the same tool that I use to organize the notes, I use a tool right now called Cursor is the name of the tool, C U R S O R. And almost everyone that uses it is a programmer. It's a pro kind of a. It's a development environment. It's basically a copy or clone of Visual Studio code by Microsoft. And so the reason I'm mentioning this is because what's interesting about that tool is that it has access to all the models. I can say I want to do OpenAI or I want to do Claude, or I want to do the newest version of Deep SEQ or I want to do whatever. And in all of those circumstances, whenever I get to the end of telling it, you know, notice what I. One of the things I said when I was describing this one is I gave it all this context, then I told it to make a new file. That tool will do that. It'll go, cool, I'll go make you a new file with a summary of everything. And now that's there. And I can use that. By the way, Claude does this as well. For anybody who uses Claude projects, it does a very similar type of thing. So when it generates the summary, now the summary of our stuff is now a thing I can use. It's a new bit of context that I can use.
- [42:58]
B
And not, not just a new thing you can use once, but you can use it forever multiple times. Right. So I think what we should do is set this up in steps so that if you're new to this, text based AI applications is the best place to start. That's how you start to learn this.
- [43:15]
A
Without a doubt process.
- [43:16]
B
So we can take whatever content you've made, whatever content you can Give it. You can take that first draft generation, like SOPs or internal documents or articles that you're writing or whatever, and you can use it. It's not a replacement for that. I couldn't just have it write me an article and submit it as is to the Wall Street Journal. The Wall Street Journal is going to be like, that's AI right. But you can use it to accelerate that content production. Once you've produced content like this podcast, we can take this podcast, which is long form, we can figure out how AI will figure out where the best places are to cut it up into short form and post it on reels. TikTok, whatever. We can do that. We can make it so that we'll. Okay, I need you to cut it down to whatever, 300 words for the caption on my social media, on my Instagram post. I usually cut it to this many characters for Twitter in a way that is going to. Obviously you would say more than that in a way that's going to resonate well. That's going to speak specifically to our target demographic that's going to get. Are you after likes? Are you after comments? Are you after engagement? Like whatever those things. Are you after acquisition? Is it a sales piece?
- [44:26]
A
That's right, yeah.
- [44:27]
B
You can take that same sort of content that you put together. You can say, help me write the email, or you can write the quick and dirty draft email, or just speak the email, transcribe instantly and then tighten it up, clean it up, get it down to this many words I've had. I do this all the time. Take what I just gave you and give it to me in a thousand words, two thousand words and I'll read it and I'll go, you know what? Give it to me in 500 words.
- [44:50]
A
Yeah, I know. I just, I do it all.
- [44:52]
B
Exactly what you just wrote was good. And 500 words with a focus on this, this and this, and it'll spit that out. So when we would used to make a piece of content, a podcast or a YouTube video or whatever the thing was, we then had to write all those things. So here's the blog post connected to the podcast or to the YouTube video. Here's the description of the YouTube video, here's the comment, the whatever the post that you make on social media, on, on Instagram. Here's the. This will do all of that for you. And you can tweak it down to literal amount of words. You can tweak small pieces. The other thing it'll often do is it'll use like emojis, but not in a way that feels like you're a junior high kid. Right, right. Like sometimes it'll use fire emoji if it's fire and it's like just enough to, to like catch your eye. Right. So that's the place to start I think is in these text LLMs, right. Editing, improving SOPs, existing content. If you already have 10, 20, 100 SOPs, you can run every one of those through AI and you should absolutely to tighten them up, to clean them up, to format them the same. So formatting is this font. You know, subheadings are in this color red. You can give it the thing like regular text is this like and boom. And now every single stop looks the same. Is corporately branded instantly. And you talk about the paid and I assume most people know this but like the paid services are cheap. We're talking about like 20 bucks. 20 bucks a month. Yeah, that like in that wheelhouse is what it is. And so you find the one Chat GPT is a great one to use. Gro's a great one to use. Claude's. Those are my top three and there's there like you said many others find one, start playing with it, have fun with it, then add another one and then I'll compare and contrast. A lot of times.
- [46:40]
A
Yeah.
- [46:40]
B
Sometimes I even say to. To grok. I asked ChatGPT the exact same thing I asked you and here's its answer. Now compare and contrast your answer to this answer and I want you to give me a combined the best of both.
- [46:55]
A
Yeah, I've done that.
- [46:56]
B
It'll spit that out and then I'll go that's better too, right?
- [46:59]
A
That's right.
- [47:00]
B
So you can do all of those things. So the thing to not get overwhelmed with because I think we can overwhelm. I think what we do here is a two part series and come back next week and talk about once you've played around with this. If you're listening to this podcast on a weekly basis, the action takeaway item is to if you're not using it and a lot of you probably already have it, use it a bunch this week. Just have fun with it. Just ask questions. I ask questions all the time. Not even about business. About like you know, you can ask questions about your like what do you know about Matt Reynolds? What do you know about the book that he wrote? What it like whatever I can I can ask. I've asked him questions about my family to see how much they know about my wife and kids and things those Things are surprisingly like small. And so I get a general idea of like how much privacy is around that. But you play with that text based application once you get comfortable with the prompting and the spit out and like, okay, this is really good. And this will take me to. This will help me put out that content in long form, mid form, short form, social media, email, et cetera. You're in a really good spot to then take that knowledge that you've built and start to turn it into. Now how do I use that knowledge to help me make decisions using AI for decision support and how to start to implement those things at the next level? Is that, what do you think about?
- [48:17]
A
I think that's exactly right. And I actually think there's a third tier and we have to decide if we'll get feedback from you all. You all tell us what you want. The third tier is you can start using it to automate a bunch of this stuff.
- [48:29]
B
That's right.
- [48:29]
A
And make it so that you don't have to even gather the context. You tell it roughly where the context is, it does the rest of the work.
- [48:38]
B
Right.
- [48:38]
A
This is the direction that so many of these things are going. And really you, you don't, you, you get value out of it in day one, the first day you do it and you start asking it questions and stuff, you're going to get higher value than you're paying for. That's, that's the bargain we want, right?
- [48:53]
B
Sure.
- [48:54]
A
But every step you take further I actually think is multiplicatively higher, right? I think it is. It is.
- [49:01]
B
There's an exponential increase, there's an exponential.
- [49:03]
A
Increase as to what happens. And it gets to the point where you will come to find that there are almost infinite possibilities with it. And I think this is where the fear starts dropping away. I think as people start utilizing this and go, is this the end of work as we know it? You start realizing it is. It is. Absolutely. If the world stays exactly the way that it is right now, that's true. Then everything that we currently do is going to get automated in some way. I'm sure we'll have robots and all those things or whatever. We won't. The world won't need us for any of those kinds of things. The thing is that's not actually what's going to happen because as these new opportunities open, there's a hundred billion new things. There's a new universe of opportunities that no one ever saw or thought of before.
- [49:47]
B
Every time technology changes, even from like the industrial revolution and then all of A sudden we make all these things. Well, you know, man, if we don't have guys to row a boat and it's a steamboat, it's a steamboat, what are we going to do with all the rowers? Yeah, well, they're going to find a different job that's a better, more productive job that often will make them even more money very quickly. And the same was the case with the Internet. And now I think we both believe this is probably the greatest change in.
- [50:13]
A
Human history for this, the greatest one I've ever seen. It's insane. Yeah.
- [50:17]
B
So absolutely the day is coming soon, in a few years where we talk about hiring somebody to mow your lawn. It's going to be hiring somebody to mow your lawn. Yeah, it's going to be buying the robot to mow. To mow the lawn. You know, buying that, which those are in the service.
- [50:32]
A
You're probably going to pay a service fee, like a somebody, the thing shows.
- [50:35]
B
Up and just mows your lawn and mows your lawn or does your dishes or make sure, like that stuff is coming. But for now, I think step one is get really familiar with the LLMs, with the, with the language models in one of these major AI, you know, chatgpt, Grok, Claude, whatever. And then once you feel like, okay, I now know how to prompt this thing and get the answers I want and get it tailored to my needs, then I think we go to step two. We go now, how do you take that and use it to help you make decisions in your business, to significantly scale faster in the business? Because again, it's not seen it through the lens you are. It doesn't have the emotions tied that you have. So now that I know X, knowing X now you can start to prompt it on how to have it help you make decisions. And then I think we can even go further. And that's. I'm at step two, you're at step three, and I think that would be a really interesting conversation. So I think that's the takeaway for this episode is just start playing around with it. You'll be really surprised at how easy it is to use. My kids use it. Your kids probably use it too. They ask questions. I mean, they're using it instead of Google at this point to ask questions. They want to know any answer to any weird, you know, whatever. Like, you know, who's singing this song or this actress that I'm looking at right now on this TV show, like she was in something that I've seen in the past. Like, you know, you don't have to go to IMDb and like try to find the thing anymore. Like, it instantly knows those things well.
- [52:04]
A
And it has the ability to capture images and it has the ability to like, in the case of ChatGPT, talk back to you. I'll give you two little mini examples of places where people that are like, I don't know if I'll ever use this. I had a tire blow this week on the side and it didn't completely blow out, it was just kind of bulging. And so I took a picture of the thing with ChatGPT on and I said, is this safe to drive on? And it was like, no, this is not safe to drive on it. Like you could see it at the Airbnb that I was at. When we shot the last podcast, the coffee maker didn't work. The morning that I woke up and I took ChatGPT and I the app and I the video, it's like live, right? It's talking to me and I'm like, hey, this thing's not working. What should I do? And it was like, oh, it's probably this thing. Do this thing. And I did it and the coffee maker started working.
- [52:45]
B
That's amazing.
- [52:46]
A
That's the coolest thing I've ever seen. So if you're the kind of person who also is like, I like to repair my own stuff or I just don't like to be frustrated in life, there's some really good tools that are out there to help.
- [52:55]
B
Absolutely. There you go. There's part one of how to utilize AI as a founder in your business to become much more effective and efficient. Save yourself time. It's not about replacing employees. It's about making employees. 1000x employees for the same amount of time or even less time than what they're working now. Same thing for you. So that the utilizing the text based AI is the best place to start. Get comfortable with it. Next week we'll take that and we'll say now. Okay, now that you've done that, now you use it to actually formulate things like the yearly game plan in your business and where you should go and what the top priorities should be in your business. Here are our biggest problems in our business that we're trying to solve. Let's do that next week, see how far we get and then maybe we'll have part three after that.
- [53:42]
A
Awesome. I love it.
- [53:43]
B
That is another episode of the build your business podcast. If you got value from this, and we hope you did, I certainly did. I actually took several notes while you were talking on my phone that I, I will immediately plug in to to my AI and and get some feedback on. So if you did get value about this, we would love a five star review on Apple Podcasts or wherever you listen to your podcast, Spotify, so on and so forth. Rankings are staying great on Apple Podcasts if you have even so I'll, I'll just give you a little tip. I listen to podcasts on Spotify. If I'm listening to audio, I watch a lot of podcasts on YouTube and because I have YouTube Premium I will often still play them on YouTube. And then because if you have premium the video doesn't have to be playing. It could be in my pocket and I'm listening to it. I can drive while YouTube is on. I'm not watching it right. But I'm subscribed on Apple Podcasts because that's where the ranking comes from. So if you haven't subscribed on Apple Podcasts or you've got friends that listen to this podcast and they haven't subscribed on Apple Podcast, even if that's not where you're listening to the podcast, subscribe on the podcast. Help us drive up those rankings. It'll help with the algorithm and we can help more people. That's really the goal here, is to help put out as much content as we can to help you up and coming founders. Certainly the day will come or we'll be talking to founders who are hopefully this includes you, those of you who are listening now, who are up and coming founders are seasoned founders. Yeah, eventually I expect the podcast to be talking to executives but right now this is a great place to start and so start playing around with with the AI models are a lot of fun. Nothing to worry about especially in the, in the text based piece. And we'll see you guys next Friday.