Ep. 5 - Thoughts about AI for small businesses - UM Live

Okay, welcome to Uncle Marketing Live. I appreciate you being here. I, uh, I've got my Superman shirt on, Superdog shirt, crypto celebrating release of a new movie. I wanted to spend today talking about AI specifically for small businesses, and this came about because I was in a session this week Where an AI consultant was addressing small businesses and how they would use AI for marketing. It was mostly valuable. They ended up opening up people's eyes, and I think a lot of small businesses still are not getting how AI is impacting their processes. So the consultant's approach was to kind of just start sharing and listing tools and services that were out there, tools for research, for content creation. It was fine, but I don't actually believe that's a strategy, and that I want to talk about why today. By the way, if you are on the chat, I see Dave from Inkshots here. Thank you for joining. If you're on the chat, feel free to ask any questions. But today we're going to focus primarily on how AI can help small businesses. Actually, it's how AI It's how small businesses really need to think about AI in terms of internal strategy within their company. Right now, there's so many tools out there. Probably by the end of this stream, there'll be 20 more launched. Part of that is because creating AI solutions right now is getting much easier. And the funding for these companies in the space is kind of off the charts. They're— everybody wants a piece of the AI pie, whether they're established players or they are new companies or they're just getting right out of college. People are building these solutions, and for a small business, it can be very daunting. You've already got a lot on your plate. As you go through your day just running the business. So now you have to contend with this monumental shift in how the entire business world is going to operate. It's a little crazy. And so I wanted to focus today on kind of how to create a mindset for dealing with this kind of input. The first is to be what I call pro-reactive. So it's like a combination of being proactive and reactive, meaning that there's a moment where you can sense a problem or you wonder if something can get done better. If you go in expecting that a solution could exist, that there, there has to be a thingy for this, you're probably right. Because in most cases, if they're big enough problems, somebody's trying to solve them. This is just the nature of the beast. It's sort of like when the internet was created and everybody suddenly had a blog, and then there were blogs everywhere about everything. This is not all that different from that phenomenon, which is, uh, if, if there you can create an AI just to do a very specific task now. So kind of the first step is just assume there might be a solution. So that means you're going out and you're going to go search for one. So you have a thing, you have a, you have a problem you're trying to solve, there's probably an AI for it. Let's go see if there is. Let's See if look for an automation solution. You can do it through search, regular internet search, Google, DuckDuckGo, Bing, whatever. Or you can actually go to the GPT in some cases and ask it if it knows of an existing tool. And chances are there is, or at least it comes close. So you'll have like some direction as to. Where to focus your attention. But using AI to filter AI tools is actually a very, very simple thing you can do just to gauge where the problems and solutions are for whatever it is you're trying to solve. You should know that not every tool that's out there is going to survive. So don't give in to kind of the FOMO for the latest and greatest thing. Every— like I said, every week there's more and more tools, and they get replaced. They lose funding, they, they can't keep up, somebody supplants them in a different way. There will always be more and more tools. One way to gauge if there is actual traction with those tools, look and see how many users they have, or if you get a sense of how many users they have, what's the buzz around that? This is where that community effect really works. By the way, there's a community workshop I'll be running July 30th. I'll post the link again at the end, but I hope you can join for that. Not all tools are going to survive. And so you kind of want to have some short-term thinking around how to solve a problem. That's okay, because chances are things get more robust as they get built out. So don't fight the urge to be everywhere, and don't necessarily fight the urge to be ahead of the game. Just exist in it knowing that You're part of the ocean of this thing that's happening. The next thing is you shouldn't be afraid to wait it out with some of the big boys. So larger institutionalized companies like a Microsoft or Adobe or your ERPs and your CRMs that are well established already have a customer base. They're going to continue to add AI capabilities into their solutions. No one's waiting for anything. I mean, they're just— they're going— they're never going hard into it. Adobe— some of the AI things you can do inside Adobe now are kind of amazing. So there's no need to throw out old tools that are there as long as they're— they have a roadmap for including AI into some of what they do. And by that being integrated into that, it's— you don't have to go out and source a new tool for that. That's actually not a bad thing either. It gets your feet wet and you continue to advance and, and take advantage of what's out there. A lot of AIs are probably going to level out. The different AI tools are going to level out and then have a point in which there's not gonna be that much difference between one thing or another. So, so whether it's an established tool inside an existing install tool versus like a brand new thing that showed up yesterday on, on Product Hunt or whatever, the, the, the ability for these tools to really be kind of amazing and requires you to take that extra step for installing something new. You need to go on a case-by-case basis, but in general, you probably don't need to watch out for them too much. Just keep an eye out on them, but you don't have to really throw out everything you're doing. Another strategy is to make sure you lean on at least one core GPT. GPT is like ChatGPT, Perplexity, something like that. Um, this is the one that you, you sit with daily, that you, that you query it daily, that you have— I don't want to say it— a relationship with. Uh, so you have a relationship with the GPT in that it knows you, it gets you, uh, and you've invested a good amount of time on it so that, uh, you've been coaching it around— whether or not you know it, you've been coaching it as to how you think. And it understands you pretty well after a few queries. And so you want to continue to use that One Core GPT as your kind of partner in all this. The next is to kind of start compartmentalizing work areas, uh, and focusing your search for solving some of those work areas that you don't necessarily want to do as much, or you'd want to get better at, and you need, um, uh, something virtual to help you get there. So that could be something like customer support or service or making reminder calls or making even sales calls potentially, that, that's where you might bring in a chat or voice assistant that's capable of doing those things. There's some— it's pretty advanced right now. It still has a little bit of a lag, but I think a lot of people just kind of appreciate how advanced they are. The voice is very natural sounding. Um, they're able to make phone calls inbound and outbound, deal with inbound calls when a customer calls in, or outbound calls, uh, to set a reminder for somebody, or just potentially even try and book a meeting. Those kind of tools exist now. The other would be— another layer of that would be an assistant that's well-trained on what the business does and can answer some basic questions, uh, and then punt it to you, the expert, so-called expert, when it can't answer something or, uh, when a deeper, more customized conversation needs to happen. They do exist now to do those things, and it's kind of powerful. It's really cool to watch. Um, Maybe you need to streamline content and design. So, uh, especially if you want to personalize something, or you just jotted down some notes and you want like a real treatise, you want something well written and clear, um, even from your notes, uh, definitely have GPTs for that. That's a great use case for them. Uh, design tools that you need to fill out things. This is kind of the magic, right? AI helps you fill out things where you don't— you have gaps, and it does it really well, especially if it has a lot of material to draw on. If you go into like Adobe Photoshop right now, you can select the image and then extend the image, uh, based on like if, if it's a sunset, now you can say Okay, make the sunset wider, add some trees, add some, um, add some— augment this image that I already have. It actually does a decent job of trying— of doing that. In fact, uh, this is crazy, um, and kind of way off topic. This sphere is showing The Wizard of Oz. So the Sphere in Las Vegas, big globey venue where we saw U2, they're actually building— they're actually going to present The Wizard of Oz. And so it's like, well, hold on, this is a movie made in the last century, almost 100 years ago. And so now you're going to show a movie that was made 100 years ago on this high-end, super high-fidelity Amazing screen. What are you gonna do? It's gonna be like this big, right? No, actually, they had it— they're engineering it in such a way that AI is filling out parts of the movie. So when you see Dorothy's house in the original movie, you're gonna see all of Kansas across the rest of the sphere. So it really kind of fills out the space and That is totally within the bounds of what AI can do because that's great source material to start from. So it just extends to its heart's content, to your heart's content. It's very detailed work, right? So it's not something that you wanna spend a lot of time on. You don't wanna spend your time drawing each tree, then using AI to handle that part of it Being able to identify incongruous data, for example, is another way to utilize these tools. Very, very helpful for that. And then eventually you'll find out what the AI is good at and what it can't do, at least temporarily. You know, you want to kind of pulse check that, but you want to get in a place where you are utilizing it because it's too valuable not to right now. It, it makes perfect sense to continue on that road. Um, there is no future before us that AI doesn't exist. So from this moment on, AI is here, and so you gotta have to deal with that. Um, you could also use AI to build your own processes, so automating processes that you have If you are currently copying and pasting anything, if you are repeating anything, if you can define a problem and envision a solution for it, there probably is a way to create software that handles that. Some good prompting on your side, uh, a little bit of knowledge on how to build a tool But it's, it's there. So I'll use an example for myself. Um, in my Apple inbox— so this is my mail email box— I wasn't doing a good job, uh, creating contacts from the people I've been sending, uh, emails to for business emails. And so, uh, they're just there as sent. Um, and so they're not there as contacts. I would basically go back into the search function in the mail application if I wanted to send back to somebody and to look them up. But the amazing thing is Macs have a scripting language. macOS has a scripting language built into its operating system, and ChatGPT knows about it. So I asked it, hey, can you look at— if I give you access to my email box, can you go and extract all the emails and contacts and create me a contact list? Um, and sure enough, it created the AppleScript for it. It asked for the right permissions. I ran it directly on the operating system, and it went into the mailbox, sucked down all the— all my sent addresses, tried to determine the names based on what was there. And, and voilà, I got myself a new contact list strictly from extracting it from email. I could have done that. Sure, I could have sat there hammering 2 years worth of emails to try and grab the contacts and go add to contact, add this contact, add this contact, create new contact. That is not a good use of my time. So instead, finding solutions like that where You know there there's probably something. Oh, by the way, I had like search for new tools first. It's like, is there anything that extracts this? No, no, no, no, no. Okay, so there weren't any new tools, but then I created an AI-based solution, or AI created a solution for me based on my needs because it knew how to get into the operating system. That's wild, and. Incredibly enabling and powerful because, you know, then I could have a more, uh, considered, um, uh, database. And you're asking like, well, why don't you just have a CRM? Sure, I could have a CRM, but, you know, I'm not— I don't need that right now. That's not— that's not what keeps me agile. The, the cost of building a CRM right now Is not, or employing a CRM right now wasn't what was letting me be agile. The fact that I, you know, 90% of my communications go out through email only gave me an opportunity to now like really kind of systematically take an approach. But at the time, I was just flying. So I didn't need, I don't need a CRM right now. It's not super critical for my kind of business, might be for yours. That's great, but for what I was doing, I didn't need it, and this was way better anyway and cheaper. So kind of wrapping things up a little bit, the important thing to remember as you're dealing with these AI tools is to remember your value as a human being. And that is to keep on learning and experimenting with what you have. You want to carve out a section of your week essentially to focus on the AI age as it applies to your business. You want to give it experiments, give it tasks, see if you get any value that way. Try things. Just— it does not get mad at you. It— you might as well just try it because it doesn't get mad, it doesn't get impatient, uh, it doesn't get rude. It is— it's merely there to accept a request and try and deliver on it. So take that as it— as an opportunity to experiment and get better. Even if it can get you halfway, it probably did it in a fraction of the time. So even though I was able to extract all my email addresses, I still had a few cleanup things I needed to do. I want to— I need to fill out the company name. By the way, I had another script that like would go in and look at the email domain and try and figure out what the company name was. It was right like 60% of the time, but for, for my needs, it meant that, oh, I just had to correct a few. I don't have to go in and be super religious about fixing everything. It just knew how to do it and took a stab at it. It's a lot like having a really enthusiastic intern. So you want to utilize that. Even if it gets you halfway, it's better than taking it— better than not using it at all. One of my other favorite ways to use GPT in my kind of weekly session with it is as a thought partner and kind of devil's advocate. So this is where I look for weaknesses in my own thinking, and I utilize the GPT as a thought partner to work through that. Tell me what— where the weak points are. Tell me why something like this won't work, and then let's work through it based on what you know And what I know and what I can introduce. And it's a great way to kind of have this process for refining your thinking and getting more concise about how you want to approach something, as well as knowing like all the tons of details that probably needs to go into it as well. So Being able to do that with somebody you don't necessarily have to pay at an hourly rate, or being able to do that and really come in strong. It's like, even if you have, like, say, a consulting agreement with somebody, you can still do this exercise, really refine how you want to spend your time. And then when you're in that consulting situation, or when you're in the customer situation, you are bang bang getting the clearest possible, uh, exchange with that because you've already kind of game planned it with AI, with GPT. And that allows you a certain amount of, uh, it makes you more confident in what you're thinking It makes it clearer to what the other parties are going to need. And really, it gives you an opportunity to kind of practice what it is you need. And so I use it like that all the time. I also use it for like, I just have some crazy business idea and then we'll workshop it within GPT. It's something— it could be a business I have no interest in actually deploying. I just think it's fun. Uh, that makes me weird, then I'm weird. But, uh, so we'll go into GPT and we'll talk through it. It's like, I think one time it was I wanted to see if I could open— what would it take to open an indoor pickleball, uh, business? And so we worked out like, oh, you need this much space for this many courts, and this kind of membership might work, but then you're gonna have to pay You're staffed this and your insurance is this. And so we really kind of worked it out that way. And then now there's another tool called Source Table, which is pretty cool as well. It is a spreadsheet and you can go in the spreadsheet and start asking it questions to start modeling like, okay, what's my cash flow going to be in 2 years if I do it this way? What if I change this? How will that affect it? What if I want to be more aggressive? What if I want to be less aggressive? Uh, how much time do I need to do this other thing? And so we're doing the calculations that way. It's pretty cool. Uh, it's— these are all things that, you know, even if you got into a situation with, uh, a consultant, having done some of this pre-work and thinking, then you can take that with to the consultant and go, you know, this is what— this is where I'm at. So you now have to put coals into it. You, you're the expert, right? Um, and that is incredibly powerful. Uh, or let's say you have a mentor relationship with somebody and you want to get the most out of that because that's— their time is valuable. You want to really utilize that time. Um, using GPT as kind of like the pre-conversation around that and pre-thinking, it really kind of helps shape what it is you need. Cool stuff there. It is AI. It's going to mess up sometimes, as this week proved with Grok. That was not a good showcase for the capabilities of AI done professionally and at that level. Kind of embarrassing, actually. Check its work. You, you know, look for kind of telltale AI signs that might be like weird corporate phrasing that doesn't really apply to something. Uh, have it tone it that down. Sometimes there's like techno bro speak that gets introduced for some reason. Uh, there was a whole incident with how to spell strawberry for a while. There's sometimes an overemphasis on some things versus a other another emphasis of the right emphasis, a right amount of emphasis. You're going to see things like that, but human beings make mistakes all the time. So don't don't stress about that part of it. Just know that you're going to have to check it. You are the editor-in-chief. You are the taste police. You are responsible as the owner of the presentation that it comes off correctly. You are, you know, the front row audience for this. So you still have to check it. But think of it this way, the time you've saved not having to do the actual grunt work means that you do have more time to validate what's going on. So don't take— don't treat it like one and done and it's out of there. Make sure you are checking, uh, whenever you're asking it to do something new. The— that's on you. That's on you and your business. So you still have to do that. And then, as you know, AI cannot really figure out taste. It doesn't really understand it. It doesn't— it's only reacting to it. And so there's, there's no— or rather, it doesn't react to it. Um, there's no doubt that AI can create some crazy amazing things. Uh, if you can imagine it and ask it, it will try and create it, and it'll be very detailed, but it itself is not an audience, so it doesn't— it's sort of— it's— it can't be your audience. Your audience is another human being, generally speaking. There's a— there's a— there's a potential that AI is also an audience because an AI is going to read your thing, summarize it, and if it came from another AI, it's going to be summarizing something that came from AI. That's all going to happen anyway, so don't worry about that part of it. Your priority probably is to focus on does this make sense for a human being? You can use it like I said for filler, but it actually has no way to know if something is good or not. It just is creating based on kind of like the past history of things, and so. That's fine for filler, but if you are trying to do something fairly disruptive, fairly unique, AI cannot make less logical choices. It can't make these leaps of creativity and courage that only a human being can make. And so that again is on you, um, to make those leaps. And take those steps. It can only go as far as you ask it to. And so it's still about you. So I will leave it at that. We're going to have a short session unless people have questions. But otherwise, going back to the workshop, That's coming July 30th. That's at 10 AM Pacific. We're going to be talking about how to use community as a complement to your brand and how to grow that community. What does a community mean in the case of business? How do you come off creating community without it being like a Shark Tank where you are just dropping people in and hoping and getting them absorbed by sales, that we're going to focus on communities that can thrive within the brand. Um, and even small businesses in particular can have an impact that way, where we will be with all this AI that's out there, with all the content that's out there The things that actually break through are going to be these kind of interpersonal relationships that exist, and you cannot manufacture that. You can only create opportunities for those things to thrive. So you're— so if you want to sign up, by all means, it's completely free and interactive. It's not going to be like me sitting and talking all hour. We're gonna actually do some fun stuff. Hopefully that, and if you're a part of it, then maybe you come out of it and you get like a full plan about how you want to take your community forward and take your brand forward through the community. So definitely sign up for that. If you're wondering what that image is, that is the top of Mount Baldy at sunset. Mount Baldy is, or Big Baldy, at sunset. It's a mountain Ridge that is part of the Kings Canyon Sequoia National Park. I was there with my family last week, which is why I didn't get this out very much— didn't get out very much for this week. But we were there. It was our standard family camp that we go to every year. We've gone for 10, 11 years. And so it's a big recharge, and obviously it's very beautiful up there. And I'm very proud of that image, not just because it's a great image, but, you know, when you make up that— when you make up to something that high on your own, then it's pretty cool. So with that, I will leave you till next week. Thanks for coming. Thanks for listening. And I appreciate all of you. Thank you.