Can MSPs Really Profit from AI? Hype vs. Revenue | EP105

[00:00:07] Hot Off The Press Dude, Ted Lasso is coming back for season four. Oh, that's awesome. Yeah. Sudeikis apparently said like he's going to do, I think if I remember what I read in the article, like season four is going to be like him coaching like an all women's team. Will that give us a reason to bring Marnie back and have her back on the podcast? Oh yeah, she's like a huge, she's like a huge fan. I mean they wrote the book. They wrote the book, yeah.

[00:00:36] Yeah, Marnie, if you're listening, drop us a Facebook comment, facebook.com slash group slash all things MSP. You're in it on whether or not you'll come back on the podcast after the Ted Lasso season four premieres in 2027. And we'll have you back on. Actually, I really think we should do like, we should have a spinoff show called the All Things Ted Lasso show.

[00:01:01] I love podcasts, by the way, that I don't really listen to enough because I'm only in the car for seven minutes a day. But I really do like podcasts that review TV shows. Like I remember there was one about Saved by the Bell. So they would like watch the episode, not during the podcast, but the podcast was them recounting what they saw in the episode.

[00:01:22] And there's like one particular episode where like Screech in quote unquote invented like this really awesome like marinara sauce. And how Max, the guy who owns the burger joint, who has like a restaurant establishment, lets the kids come in and use stuff that they stole from the science lab in their high school to bottle Screech's marinara sauce and film a commercial there. Meanwhile, there's like, they're breaking like health laws. This is what they're talking about in the podcast.

[00:01:51] Like they're breaking health laws and there's definitely like commercial tax implications to all of this. Only to find out at the end of the day that the marinara sauce was like his grandmother, Screech's grandmother recipe that he read on the back of like a Campbell's can. And they had to shut it all down. Like I digress. But yeah, Ted Lasso is coming back for season four. Marnie Stockman is going to come back and talk all about it for you. I've, I've, I've, I've deemed she's coming back.

[00:02:19] That's my word is my, my word is bond. Ben and here, near, near, near, near. This episode is sponsored by navigating compliance requirements should not be a headache with compliance scorecard. You can track report and manage all your clients compliance needs in one place.

[00:02:43] You can stay ahead of audits, ensure client satisfaction and reduce risks with ease. Start simplifying your compliance today at atmsp.link forward slash compliance scorecard. Simplify IT procurement with Griffin IT. They provide MSPs with access to equipment, pricing and product availability not typically found in traditional distribution channels.

[00:03:09] From refurbished electronics to IT asset disposition, Griffin IT offers solutions that boost margins and solve supply chain challenges. Visit atmsp.link forward slash Griffin IT to learn more about how they can help your MSP succeed. What's up everybody? Welcome to the All Things MSP Podcast. I'm your host, Justin Escar. With me always is my good friend, pirate, producer extraordinaire, Mr. Eric Anthony. Eric, did you know that apparently fish cough? That's what my Snapple cap says.

[00:03:38] Real fact number 38. I don't think these are real. I think these are AI. Oh, well, you do know that 53% of MSPs polled say there's going to be a significant increase in revenue due to AI specific services. Ladies and gentlemen, that's what we call in the industry a layup.

[00:04:04] I actually, by the way, I almost did this, which I've mentioned the Basement Yard podcast before. There's an episode where one of the guys goes, lay up, and the partner goes, that was a free throw. I had to consciously think of what a layup looks like. Yeah. Look, we've talked about AI planning on this show, and everybody knows our hot takes on AI and where we're going with things.

[00:04:31] And so we're going to talk about it again, because clearly we can't get enough of this, much like people can't get enough of dark mode on websites. It's one of those things that they just stick everywhere. Let's look at history, right? Bakers are like, we're going to have bacon and everything. It's the same thing. We'll put AI in everything. Okay. So let's first define two things. Let's define, we have generative AI, ChatGBT, DALI, MidJourney.

[00:05:00] And then we have like machine learning automations that are now just being called AI. And so I want to differentiate those because I think there's a lot to be said about both. To Eric's point, though, it's too generic. And we're probably going to end up just muddying the waters again. Well, and I think you have to include the third type of AI that is just now rearing its ugly head, if you want to call it that.

[00:05:28] And that is agentic AI. And I say rear ugly head for a reason. And that is because people are treating these AI agents as if they're people. They're actually creating teams of these AI agents to do specific jobs. So there's a lot that goes into this.

[00:05:51] Of course, the numbers that I just read to you come from the new GTIA, formerly known as CompTIA, the new Global Technology Industry Association State of the Channel 2025 report. And if you don't have a copy of this report, I will link to it in, you know, whatever description post, et cetera, that you're seeing this from. A lot of great information.

[00:06:17] But this is one where Justin and I thought we definitely had some opinions and that hopefully we can give you some direction in terms of how much of this is hype, how much of this is real. And what is the actual opportunities, if any, to generate excess revenue for your MSP?

[00:06:46] Take it away, man. Like my opinion on AI. No, I'm just kidding. Right? Like my thing about AI is that like, I mean, I said it earlier. I was like half kidding, but like I feel like we're just, it's being jammed down everybody's throats, right? Like Google and Microsoft are now upping their pricing because they're enforcing Copilot and Gemini, or reverse that, in to their product.

[00:07:14] And like I've tried Gemini and so far it's just a Google search. And I swear that Copilot is just Clippy in disguise, right? Because if I open up Word and it's like, hey, I think you're trying to write a letter. You want help with that? I know who you are, Clips. Get out of my face, right?

[00:07:33] Like there's, to me, the problem with AI is that it got super hyped, super fast, and everybody thinks that they can use it without any sort of recourse. There's, there's, there's, there's, there's, there's a problem when you just let everyone loose on it. And, you know, when you're going to have a business, i.e. not just at the IMSPs, but like our clients, they have a business they're on.

[00:08:02] They have a brand. They have a name out there. And then they're going to go and like muddy that by just letting, by just blindly letting AI write content or, you know, make images for them. And don't get me wrong. Like I am just as, I'm calling it kettle black here because like I literally, before we started recording, I was like writing a proposal for someone, like a quick email. And I threw it into JadGPT and I was like, make me sound smarter.

[00:08:32] But like I read it first before I copied and pasted. But so many people don't. They just like blindly copy and paste it across. There's times to use it. There's times to not. But I think really what all of this comes down to is we as technology experts need to help our clients write policy around AI usage. Whether it's any of the three we mentioned earlier. It's about policy and it's about writing.

[00:09:02] The funny thing is you can use AI to write that policy. Yeah, except I don't know. Like that is a little bit dangerous, right? Because what if AI tries to rewrite the policy so that AI wins in the end? So yeah, be careful of that.

[00:09:24] But I think that's a great example, Justin, of one of the ways that MSPs can legitimately monetize how their clients are using AI. It's providing that guidance and those policies on how their employees should be using and should not be using AI. Because there are risks, right?

[00:09:50] There are risks to putting your private data into a public LLM. It's not good. And so there are things that you need to do as an MSP to help your clients avoid those problems. So, you know, that's like I was saying, that's just one example, right? Of things that an MSP can do to help their customers and be able to charge for it, right?

[00:10:19] Because this is also not like many policies, is not a one and done. Because AI is one of the most rapidly changing things in our industry, right? And also, I think that the MSP, if you truly are a managed service provider, and I know that Justin and I both have very strong opinions on what that means and what it doesn't mean.

[00:10:43] If you're truly managing their technology, you need to be staying on top of this and helping them navigate not only what they should and shouldn't be doing, but what is hype and what's not hype. Because there are legit things that if they're in a certain industry, they can make use of in terms of generative AI or even now with agentic AI.

[00:11:14] But it's very difficult for them because everybody and their uncle is trying to sell them something that is augmented with AI. Yeah. The question is, is it actually useful to their business? The better question is, is it really AI or just an automation, right? True enough. This presentation I was doing, there's a guy on who's talking about he's from a cyber firm or whatever.

[00:11:43] And he was very dead set on trying to separate the idea that AI is just a better version of automation, right? And while generative AI like ChatGVT does help pull information together, it's just a really, in reality, it's just a better search engine, right? And so the agentic AIs, like there's one I signed up for called Cintra.

[00:12:14] And it was like, it was supposed to be like, they're really cute. I fell into their marketing a little bit. But like these cute little astronauts. And there was one that was doing SEO. And there was one that was doing biz dev. And there was one that was doing sales. And I tried all of them. And they were all just equally horrible.

[00:12:31] Like every, I think the problem is that everyone is on that bandwagon and they're riding the hype of AI and no one really knows the direction we're going. Like we're on a train that has got 70,000 different tracks that it can go on at any given time. And no one knows which one we're going to take.

[00:12:55] So if you whisk away a lot of like the little ones and let's just like focus on like, you know, ChatGVT is obviously the forefront and DeepSeq, the one from China, which, you know, don't use that. But like they're at the top of the line of it. But again, it's one of those things where like, yeah, you know, we joke around. Like don't give your data to DeepSeq because it's in China. But like don't give your data to ChatGVT either.

[00:13:24] There's so many people who have personal ChatGVT accounts and they just will just like blindly just upload information. And they shouldn't. I was trying to read the ChatGVT like rights and in terms of service about how to do things like what's collected and what's not.

[00:13:46] Because I was always under the impression and maybe I'm wrong that ChatGVT for business will not retain information the way ChatGVT for personal use. Like if you use it on the free or the personal one, like you upload data, you're feeding the source.

[00:14:07] But if you do it under business, like under Teams, I want to say that I read somewhere and I could be 100% wrong that your information in that regards will not feed into the data. But maybe I'm wrong. The other thing is, you know what I learned about recently is Sora, the new video generation one.

[00:14:32] Like we talked about this before where like there was that dude who was on a call with seven people from his quote unquote work. Right. They were all AI and he got somehow or other like he was like a financial officer and he like transmitted like $10 million, some ridiculous amount of money out of the business into like some Chinese bank account. Like because the AI told him to like these horse, there's good stories, right?

[00:15:00] Like the doctor who put in all of the symptoms and got like this acute problem figured out for this patient, but only to find out that the sources were all made up. Like there are stories like that all the time. These are the things as the IT experts that we should be sharing these stories with our customers when they come to us and say, can we use AI? Now, I've lucked out in our industry and the type of clients I have.

[00:15:26] I have like one or two customers right now who are like knocking on our door about AI, right? Most of my customers are not asking about it, which has been saving me a lot of face. But I know a lot of MSPs are really struggling with this because what should they be allowed? You know, what should the customer be allowed to use? And should you as the IT expert tell them, hey, maybe that's a bad idea?

[00:15:52] Actually, I was on this call and there was an IT person on there and he provides secondary support to a very large, apparently the guy was telling me, he's like a very large industrial manufacturer and they have internal IT.

[00:16:07] And internal IT emailed him and he said, hey, how can we use AI to basically, he was asking like, how can we use AI to check permissions in Azure to ensure that people have done X, Y, and Z within their account and have XYZ permissions and have XYZ settings or whatever it is. And he's like, I don't even know how to write back to this.

[00:16:33] Like my aunt, like I was like, tell them not to be idiots, but because like, that's not a thing. And also if that were a thing, you're basically working yourself out of a job. Well, I mean, but people are trying to do things like that, right? They're trying to take, I'll call it unstructured data, right?

[00:16:58] And through the use of AI, get real answers about behaviors or statuses or timelines based on a group of documents.

[00:17:14] My problem is, is I think that we've shown so much that even if AI is not blatantly lying to us, which is a conspiracy theory held by some, it definitely hallucinates.

[00:17:33] And so just like asking a C plus student to summarize war and peace is chat GPT. I shouldn't pick on chat GPT is AI going to do the same thing.

[00:17:52] That's why I don't know that asking it questions about actionable data. And I guess the way I want to describe this is, you know, the example you used, they wanted to go out and look at Azure, look at configurations, look to see if people have done certain things with certain files or whatever. People are going to be measured.

[00:18:22] They're going to be graded on those things. And if the AI doesn't tell the truth or is not accurate, that's bad. Yeah. It's even worse when we talk about configurations, right?

[00:18:38] If we're talking about security configurations across Azure, if we ask AI to give us a summary of the configurations and tell us if there are any risks in those configurations, but it doesn't tell us the truth, that's an even bigger problem. Trust but verify. Trust but verify. Trust but verify, as my good mentor Craig says all the time. But you're right.

[00:19:08] Like, look, generative AI specifically generates content, generates stories, generates information. We've already seen that. The doctor, the sources were all made up. There was that one, but the lawyer who used ChatGPT and actually everything that that lawyer said ended up being false because none of the cases or the precedent that the ChatGPT had come up with were real.

[00:19:37] And the downside is that ChatGPT is the leader in this format, which means anything else is under that. You're going to run into problems. And so to your point, if an AI system doesn't know something, it's never going to say, I don't know. It's just going to cover up its mistakes. It's just going to come up with something. It wants because its job is to make you happy.

[00:20:03] And so if the system, you know, if you ask it, hey, are the permissions correct here? And it doesn't know, it's not going to say, I don't know. It's just going to say, yes, they are. And it's just here to make you happy, which is why, you know, I think there's a massive problem around it.

[00:20:24] I think right now, as it stands, when we're recording this March of 2025, our job as MSPs is to guide the customer through the journey. Be the voice of reason, be the voice of trust and say to them, listen, these are the stories that have come out in the news about ChatGPT, AI, whatever.

[00:20:55] Do you want to risk your business doing this? What is your comfortability of risk? And where do we want to allow your staff to do something and not something else? And so you can write this policy. And again, we've talked about cybersecurity in the past a million times, where it goes people, process, technology. The process is what I'm talking about here, which is setting these rules, making an acceptable, like everyone has a technology.

[00:21:24] And they're acceptable use handbook that says the company owns the computers. We can do whatever we want. Stop looking at porno on the work computers. Like there should be an entire section of acceptable use of technology as it refers to AI. And it should be updated every six months. And you know what? That's a pain in the butt. But that's something you could charge your customers for 100%.

[00:21:49] If you're an MSP right now, listening to this show the week it comes out, one, Eric needs some help fixing his internet. Two, just think about the AI conversation. And stop believing. The other thing that really bothers me about AI right now is that there's too many people out there that are trying to be AI theorists and trying to like push AI down other people's throats and not enough people that are actually like AI experts because it's too new. Stop listening to all of them.

[00:22:18] Come up with your own opinion about this. Read the articles. And then help guide your customers the right way down a path that is safe for them and acceptable by you. I totally agree. And yeah, I think we should end this episode before I just totally get attacked by AI and it completely removes my internet.

[00:22:42] This is one of those things where like you got to like you have to green screen in like the T-1000 coming up from behind you and just like right through your head. Well, if you want to see that happen, check out youtube.com slash at all things MSP. Watch Eric get plunked by a T-1000. Check us out facebook.com slash group slash all things MSP. Join the ever growing group. Ask questions in there. That's what we're here for, for helping you as an MSP. Check us out on all your favorite podcasting tools. Leave a review. Say hi. Let us know you're here.

[00:23:11] Let us know what you're doing with AI. Love to hear what your horror stories are around AI and customers. So leave that in the Facebook group, on LinkedIn, or whatever. If, all right, what's your line? Like if Facebook's not your thing, check us out at LinkedIn. We're on, we're on LinkedIn. Who knew? But yeah, that's it. Anything else? AI, Anthony? No, because I don't know when they're coming to get me. I'm going to end the show with 1-0-0-1-1-0-0-0. Bye.

[00:23:42] Hey folks, real quick. I think Eric's internet just died. Or mine. It's hard to tell. So here we are. We're just going to probably cut all this together. I'm going to send him a text. You just died. I'm still here. So what was I saying? I'm the best host you've ever had. I don't know. Something bigger going on. Sunspots. Sunspots. It's sunspots, Justin. Solar flares.

[00:24:12] The AI is coming to get you. Thank you for listening or watching the All Things MSP podcast. If you liked this episode, go ahead and give us a thumbs up. Hit that like button and consider subscribing to catch all our weekly episodes. And from your host extraordinaire, Justin Escar, and myself, Eric Anthony, your humble producer and All Things MSP founder, thank you very much for spending your time with us.

[00:24:40] If you are not aware, All Things MSP started as a Facebook group and now supports over 6,000 members. We also have a LinkedIn page for those of you who don't do Facebook. And make sure to check out our YouTube channel for even more content. A special thank you to our elite sponsor, CoreView, helping you manage your Microsoft 365 tenants instead of them managing you. Thank you to our premier sponsors, EasyDMark, Helped, Gozinta, Movebot, and SuperOps.

[00:25:09] And thank you to the rest of our sponsors. Without sponsors, we could not do what we do for the MSP community. Please consider checking them out. The All Things MSP podcast is a BizPow LLC production. The views and opinions of the hosts and guests are their own and do not reflect the thoughts and opinions of any employer, vendor, sponsor, or random taxi driver in the Metro DC area. Be sure to join us next week for another exciting episode. We'll see you next week. Bye. Bye. Bye.