Episode 229: SPECIAL: Using AI in your MSP's marketing

[00:00:00] Speaker A: Fresh every Tuesday for msps around the world. Around the world. This is Paul Green's MSP marketing podcast. [00:00:09] Speaker B: Hello and happy Easter to you. Welcome to a very special episode of the podcast just for Easter 2024. We have dumped our usual format because this week I have a very special guest on. He's kind of like the marketing r and D department of the entire channel. The guest I'm going to introduce you to in a second is constantly trying new things. I learn from him all the time. I consider him to be a very good friend, and it's always worth having a conversation with him, whether that's on email or WhatsApp or doing something like this. Because every time I'm scribbling down things, he says and thinking, I like that. That's a really good idea. I'll take that away. [00:00:46] Speaker A: Paul Green's MSD marketing podcast special. [00:00:52] Speaker B: So we're going to talk about in the next 20 minutes or so some very cool AI marketing tools that we have seen sort of being launched over the last couple of weeks. We're going to talk about something called GPTs, which is something that's come out of chat, GPT, and how you could use that in your MSP. It's far too many acronyms there. And we're also going to talk about why you would launch a brand new PSA in 2024. What's that I hear you say? Do we need another one? Well, maybe we do. We will find out later on in this special episode. So let me introduce you to this week's special guest. [00:01:27] Speaker C: Hi, I'm Kevin Kloon. I'm a co founder of MSP Growth Hacks and recently of Zest, a new PSA. [00:01:35] Speaker B: That brand new PSA we are going to talk about later. So welcome back onto the show, Kevin. It's great to have you here. I think this is your third or fourth appearance, which is pretty good. At some point, you're going to become the most appeared upon person on the podcast. Is that even a thing? Did I say that properly? It's a thing. Now we'll get you a trophy. So tell us a little bit about you and your background. Assume that whoever's listening to this right now or watching us on YouTube doesn't know who you are, where have you come from and what do you do? [00:02:02] Speaker C: Yeah, so I was an MSP, I guess back in 2018, we had an exit, and after the exit, we just sort of figured, trying to figure out what do we do now? And so I thought, hey, why don't we just start a blog and just start sort of dumping everything that we learned. And so that's how MSP growth hacks was born. We've been writing articles pretty much since then. So we're coming on, I guess five years or so of this journey, and it's kind of taking us in a lot of different directions, working directly with msps, whether it's on marketing or growth. More recently, I think we're getting more into software and solutions that can directly benefit in a far more scalable way. But over the years, like you said, we pretty much have stayed on the forefront of just like what's new, what's different, and what changes around the corner that maybe people aren't paying attention to yet. [00:03:03] Speaker B: Yeah, and I love getting your newsletter. I have noticed they're not as consistent as they used to be. Just scolding you there with that one, Kevin. But whenever you get marketing has to be consistent, right? [00:03:13] Speaker C: Of course. [00:03:13] Speaker B: Whenever I get your newsletter, I always keep it. I'm actually subscribed twice. I'm subscribed in my work email and my home email because one of them, I want to make sure I just have some time to sit and absorb and read what you've written. And as I said at the beginning, it feels like you're the marketing r D for the entire channel. So you can't stop this. You have to keep going with this, even if your new PSA takes off for our audience, if you haven't been to MSP, it's mSpgrowthacks.com, isn't it? [00:03:40] Speaker C: Correct? [00:03:40] Speaker B: Yeah. So if you haven't, if it's a website you haven't been to, I mean, I kid you, is as Kevin says, there's five years of content there. But this is not just one of those blogs where someone's filling up the blog for the sake of filling up the blog. Right? Kevin's not doing this for Google. Kevin's doing this with the content is insanely good. And I know a lot of very, very smart people who subscribe to it for free, it doesn't cost anything, but who subscribe to it and listen to what you have to say. And that's what's made you a pretty influential person, Kevin. So thank you. On behalf of the entire channel, I'm going to thank you for MSP Growth hacks. Now let's talk about AI tools. And before we do, I have to give a caveat that we are recording this in the middle of February, which I know seems kind of insane when you consider where Easter sits this year. But we work really far ahead on this show. I don't like to get behind. The reason I give you that caveat is because Kevin and I are about to talk about AI, and as we all know, you can talk about AI on Monday, and then you wake up on Tuesday and it has completely changed again. So if we talk about anything now and you think, oh, everyone knows about that, or, oh, that's been superseded by this, you now know that actually, we recorded this in the distant past. So let's start by talking about cool AI marketing tools. Kevin, I'm going to start with one, if that's okay, because we've just started testing one actually within our business, and I'm going to give you very early, like a day zero five review of it. So it's a very cool tool called Voxia. V o Xia AI. I think it's voxia. If you just Google Voxia, it'll come up, and it's come out of Israel. And I saw it written about in a marketing newsletter about a month ago, and I reached out to them and said, we want to be not quite a beta tester, but we want to be one of your early clients. And I think we're like client number seven or something. But what it is is it's an AI automated outbound calling. So we have, I wonder if I should give her name now. I probably shouldn't, but it's called an agent. It's an AI agent, and her name is Laurie. And we are feeding in numbers of msps around the world, and Laurie is calling those msps on our behalf. And what she's trying to do is she's trying to book a 15 minutes zoom with the human who works for me called Ben. That's her goal. So we have worked with Voxia to set up an entire script. It's clever, it's AI, so it listens to what you say. So if it asks you a question and you go off track, or you say, sorry, what was that? It picks that up and it reacts like a human does, it'll say, oh, I'm sorry I asked you, and it's really scarily good. So I've tested it on a number of people and said, oh, I've got a new telesalesperson. Can I get them to call you? And haven't told them it was AI. And apart from there being a kind of a slight gap in the conversation, the conversation doesn't quite flow so well. It's really good. And the day it all changed and the day that we committed was the day they integrated it with calendly. So when they first launched their first version, where basically the agent with whatever voice or saying whatever you wanted, she would say, oh, I'm going to get you booked in. I'm going to text you, I'm going to send you over a live calendar, which as we all know, very few people ever then actually booked in. But as we're recording this here in mid February, like a week ago, they finally got calendar integration, and that's been a complete game changer. So we are in very early days testing this right now. I really hope it works. I'm not going to talk about pricing because that's not unfair to them because they're trying to find out their pricing model. But it's not cheap, it's not expensive, it's kind of somewhere in the middle. And as I say, fingers crossed, if this works for us, this could actually be a really good solution for msps because it's exactly the use case. The use case for us is the same as it is for msps, which is calling people to find out if now is a good time to talk and if it is booking a 15 minutes appointment with a human. So that's my cool AI marketing tool. What's something cool that you've come across recently? [00:07:29] Speaker C: Kevin in yeah, so just to add to that, I have seen some of these outbound prospecting tools, especially the sort of outbound calling tools I know in the US, the FCC is already jumping all over this, and I think there's going to be regulation very quickly on because when you start to stack robocalling, which was already a problem now with like, oh, and the computers can talk now, that takes it to a whole different level. And this is the problem, right, with AI is the advancements are huge and abundant, but then everybody's scrambling on the back end to sort of clean up the messes that it's creating. So you can also burn a lot of time. I've been experimenting with all kinds of tools that I'm finding, but I guess recently I keep coming back to sort of just the basics of chat, GPT, copilot, custom gpts, and what I've been doing recently, I published an article last week on essentially like creating a case study, right? And so in that article, we walked through the process of, okay, let's publish a questionnaire that you can fill out for a case study and then a prompt that you can use to generate the case study and then a design brief that you can then take once you have your case study and get it designed. And so what that really is, is it's a turnkey process that regardless of whether you want to use chat GBT, you want to use copilot, you want to use Gemini, whatever, something else, you can just kind of apply that. And so this is what I've, after all these kind of experiments, and what I keep coming back to is how can we really build more so, like, processes around this to where I can repeat a result over and over? And it's not going to just all come crumbling down in the next few weeks or months when a tool changes. Maybe that's not the popular tool this is, or they lost funding, you know what I mean? It's going to start to really be a rat race as things pick up. And I think most of these things will fail, and not because they're not incredible, but just because the race is on and things will change over time. So I think that's what, as exciting as it all has been, I think what's exciting to me now more recently is being able to repeat results just using the know. And so that's kind of what our focus has been on recently. [00:10:14] Speaker B: Okay, so let's talk about custom gpTs. And before we do, I'm hereby officially making you, Kevin, our FCC monitor. Because obviously my business is based in the UK, but we're calling to the US. So if the FCC puts regulation on this, you personally have to WhatsApp me and tell me just to switch off voxia. Do we have that agreement in public right now? That's good. I don't want to get fined, right? I don't want to break any laws. So let's talk about custom gpTs. So can you explain, let's not assume our audience knows everything. I know lots of msps are very up to date on this stuff, but there are also sometimes people I speak to aren't. So can you first of all, just explain what a custom GPT is? And then let's talk about some of the experiments you've been doing with it recently. [00:10:58] Speaker C: Yeah, so GPT stands for generative, pre trained transformer, which doesn't really tell you anything about what it actually know. The way that I like to explain it is basically just, it's a version of, let's say, Chat GPT that you can essentially upload your own data to, and you can also change the behavior of it. So you can make it answer things a certain way, or you can have it sort of ask you questions and do sort of like a role playing, and however you essentially build it, that's sort of how it behaves. And so the GPT that we created is called MSP growth bot. And essentially what we did is we took all of our years of content, whether it's articles, videos, we have a bunch of PDF guides that we've created. My book, the MSP growth funnel that I published back in 2020. So everything's in there. And now, since the data set is limited, you can have conversations with it and you know that you're getting answers that are sourced from genuine experience, right? So this is sort of the problem with chat GBT, sort of proper, I guess, or in general, is that when you ask it a question related to a very specific topic, you're going to get like a million different ideas and this and that, like every possible outcome. But every possible outcome isn't always that useful. Sometimes you just need to know, like, okay, this is the outcome that's most likely based on a limited amount of ideas and sort of execution steps. So that's, I think, difference right now for us. So we encourage all of our readers to use MSP growthpot when they're trying to kind of ask questions related to growing their business. Because essentially what it does is it looks at our content, it formulates responses, and it really does a great job of stitching these things together. It might be some knowledge from three different articles, which at their surface have very little relevance to each other, but it's able to give you whole answers that really kind of COVID it in a way that personally we never could. And so currently you can visit the GPT store and on Chat GPT, and you can see all of these different GPTs that are now really useful. And so essentially, to boil it down, it's just a Chat GPT. That's for a very specific use case. [00:13:34] Speaker B: So just before we talk about how an MSP could use a GPT, how do we access your growthbot? [00:13:40] Speaker C: So currently you can just go to msptowax.com, go to tools, and then you'll see the link to the GPT. There's also a banner on the side. [00:13:48] Speaker B: Okay, really simple. And I love that. In fact, I could imagine that growthbot, of course, makes connections and remembers things that you can't because your inefficient organic brain forgets the article that you wrote three and a half years ago about XYZ subject, and perhaps wouldn't make that mental link immediately, whereas of course the software would, because it's holding all of that information. So I think that's a really smart way of using that. So the tough question then is how would an MSP use a GPT? What are some of the use cases that you think are either going to become standard or are certainly worth exploring, even if there are some risks attached to that? [00:14:26] Speaker C: Yeah. So there is obviously a lot that you can do with training your data. Right. But again, there's a lot of sensitivities there. So it's not sure if that's a great idea at this moment. But what I think where this is leaning more so is for a behavioral thing. So a good example is I actually helped an MSP who has an outbound sales team, and he wanted to basically train his salespeople on their script. Right. So they have all these scripts, they've got all this documentation, and this was actually provided to them from an outsourced partner that they have that is providing these materials. And so they were actually able to create, like, a little internal GPT that is really just as essentially like a copilot for their salespeople. So if they're getting a certain response now, they want to say, like, oh, how should I answer this question? Or what should I do? So that's a really simple sort of sales and marketing focus. Doesn't really require proprietary data that would have any sort of harmful impact, but it's just a way that you can use the GPT in a simple way that can have some sort of benefit long term and that you can build on as you learn. And as that becomes more useful, you can train it with more and more data. So this is like, what's great about our GPT is that every time I publish a new article, it then gets loaded into the knowledge base. So, like that article that I mentioned that we published last week on case studies, now that GPT has just gotten smarter, and it's just because every time it provides an answer, it's doing a live look at the content that we have. [00:16:17] Speaker B: Yeah, got it. That makes sense. And when you were talking about content that I can't remember the exact same thing you said, but you said about it being harmful. What you were essentially talking about is, how should we put this politely? If the data gets out of the GPT in some way, it doesn't matter if that's just sales queries, is that what you're saying? Versus you wouldn't put. So you could do that with technical data, for example, but you wouldn't do it with client technical data? Is that the kind of thing that you meant when you talked about that kind of data. [00:16:49] Speaker C: Yeah, exactly. And basically what's happening recently is that even though they have chat GBT, for example, has sort of like the team and enterprise versions, it seems like there's a lack of control over what information gets leaked after it's been uploaded and where it's getting leaked to. And so that's why it's sort of best practice at this point. Even though two weeks ago this wasn't the case, is that really no proprietary data related to any sort of technical or customer information should really be uploaded at this point. There might be ways that you could do that locally with a spreadsheet and, like copilot or something. I'm not really that familiar with that, but for sure, I think that currently in today's state, possibly will change. Is maybe not the safest thing. [00:17:42] Speaker B: No, that makes perfect sense. Well, let's be honest, any data safety, if there's any hint of your data not being safe, you've obviously got to err on the side of caution. I think that's a default for absolutely everyone right now. It's really exciting, and we're going to wrap up talking about AI and just briefly talk about your new PSA. It's really exciting for me at this stage to see how Chat GPT is evolving so quickly, and then we're starting to get. And you'll tell me what this is called, because I can't remember, but I've preordered a rabbit r one, which is the gadget that will actually do things for you. So Chat GPT is just organizing information. Can you remember what that's called? The overall descriptor of action? [00:18:23] Speaker C: I believe it's RPA. Right. Robotic process automation is what it uses. It remembers sort of like paths. [00:18:29] Speaker B: Yeah, that's it. And that's the next step, isn't it? So it's one thing to have your information organized and cross referenced and search and predicted and generated, and then the next level is doing it for you. And I think this is. We're almost getting to that point of being. Do you remember when Siri came in in 1842, or whatever it was? It seems like that long ago. And everyone was like, oh, Siri is going to be great. And then you launched Siri, and it couldn't do anything. It couldn't even remember what it had just said to you on the previous answer. And so very few of us use Siri. And of course, Alexa came along, was a bit better, a lot better, but it's still not exactly the same. And it feels like we're right on the precipice of that point where not that far in the future we will have a safe way to put proprietary data into some kind of custom AI engine. And for an MSP to be able to take 1015 years of PSA data and have an AI analyze that and suggest answers and look for connections and look for problems and look for upsells, and I find that all of that exciting. And then actually maybe even trigger email campaigns automatically or notify salespeople. And the marketing opportunities of this alone, they just blow my mind and it feels like we're on minute one of hour, one of day one, and we've got obviously potentially years and decades of development to come on that. Anyway, that's all of AI. I know you and I could talk about that for hours and hours, but I do want to talk about your new PSA before we finish our Easter special. So there are lots of psas around. I guess this is a question you hear a lot, Kevin, why do we need a new PSA? Tell us what you've, you know, I. [00:20:10] Speaker C: Don'T want to take too much of the know. One of the benefits of MSP growth hacks is we tend to attract some really sort of innovative and forward thinking msps. And these are msps that sort of become part of our community. And I love talking to them and helping them really in any way. And so I guess it was two and a half years ago or so, one of the msps called Lime tree labs out of Boise, Idaho, reached out and they had a really great. I had already known them and they had read my book, but they had a really interesting idea for a PSA. And the idea was really simple. It was that the PSA is too complicated. Everybody's trying to push out the ten and under crowd, meaning like ten technicians or fewer. There's a ton of barriers for that right now is like onboarding fees, there's minimum license counts, there's giant contracts, you name it. And so what if we just had a PSA that was ready to go out of the box? It was built for an MSP specifically. It required very little customization, and you could use it with essentially like no contract and no minimum seat requirements. So from the beginning, I sort of helped them through a wireframe, just sort of thinking around what they had, but they really didn't need much of my help at all. They went out and built this thing. And then once I saw it, I'd say a year and a half ago, I was really excited to see how I could get involved. So I became an advisor and a shareholder and now I've sort of been helping in more capacity as a co founder. And so it's called Zest. Zestmsp.com. And the point of this PSA is know it's simple, it's simplicity. It's about just getting up and running as fast as possible. If you have a huge team and you require all these automations and complexities and whatever, there's far better tools for you. But if you're just looking for a simple way to operate your cervix desk ready to go, you don't want to spend a ton of time tinkering because it's already built for you and how you should be running an MSP, then I think zest is the way to go. [00:22:35] Speaker D: Coming up next week hi, I'm Christy Mitchell. Come join me on Paul's podcast, where we're going to talk all about strategic content marketing for msps to help you attract, nurture and convert your ideal clients. [00:22:48] Speaker B: I'm going to set you a challenge next week. Every time you sell a project or a piece of hardware, how can you attach a stream of monthly recurring revenue to it? I've also got for you five quick wins to instantly improve your MSP's net profit margin. Join me next Tuesday and have a very profitable week in your MSP. [00:23:09] Speaker A: Made in the UK for msps around the world Paul Green's MSP marketing podcast.