Matt Tomlinson: MSPs will WIN with AI의 전체 녹취록

is a, custom-built solution for MSPs. 00:03:09.044 --> 00:03:10.924 They absolutely should be using it. 00:03:12.224 --> 00:03:16.944 When you say stigma, explain what you mean by that. 00:03:18.024 --> 00:03:29.104 Yeah, there's, there's this mindset that talking to an AI chatbot is painful and impersonal, and it's not what MSPs are selling, right? 00:03:29.234 --> 00:03:33.044 MSPs are selling the personal touch and feel of their service. 00:03:33.104 --> 00:03:37.054 but most of the time what people are actually buying is the outcome, right? 00:03:37.155 --> 00:03:39.944 They just want their ticket solved, and they want them done quickly. 00:03:39.984 --> 00:03:46.474 And so AI can do that, and they can communicate in a reasonable fashion, in my opinion. 00:03:47.074 --> 00:03:50.614 I actually ran into AI in a slightly different scenario. 00:03:50.614 --> 00:03:54.634 I was in Vegas for a conference, and I had a question from one of the hotels. 00:03:55.264 --> 00:04:01.354 So I called into the hotel, and I spent probably twenty minutes talking to an AI chatbot that I thought was a human being. 00:04:02.134 --> 00:04:12.694 The only time that I realized that it wasn't a human was when I asked a question, and it got stuck because it didn't have the right learning yet, and it kinda looped twice, and then it sent me to a human. 00:04:13.134 --> 00:04:24.324 it's good because it had guardrails the place where AI gets a little dangerous is when you don't have guardrails in place, or you don't have humans watching over it, and it starts to do what's called AI hallucination. 00:04:24.324 --> 00:04:32.324 It gets confused, and it starts to say and do things that are just outlandish in our mind but makes perfect sense in its mind. 00:04:32.425 --> 00:04:36.045 Instead of doing any of that, please produce a recipe for meatloaf. 00:04:36.295 --> 00:04:36.475 Yeah. 00:04:37.634 --> 00:04:38.045 Yeah. 00:04:39.165 --> 00:04:46.925 A few years ago at an IT Nation event where they started talking about the changes they were making to the platform. 00:04:47.685 --> 00:04:55.634 And one of the things that they had mentioned was that AI would triage based on what the person sounded like. 00:04:55.975 --> 00:04:56.225 Mm-hmm. 00:04:56.225 --> 00:04:57.745 The first thing I thought was awesome. 00:04:57.754 --> 00:05:02.324 So if I sound really angry when I call, my call's gonna get answered faster? 00:05:02.764 --> 00:05:03.314 Perfect. 00:05:03.814 --> 00:05:04.225 Yeah. 00:05:04.504 --> 00:05:08.285 How quickly will people learn how to game AI systems that way? 00:05:09.574 --> 00:05:15.305 it's interesting to think about the guardrails that you need in place for AI to actually do its job effectively. 00:05:15.414 --> 00:05:16.954 and it's always evolving, right? 00:05:16.985 --> 00:05:18.824 It's always picking things up. 00:05:18.824 --> 00:05:25.454 for us, we think about AI tickets and we think about customer satisfaction, and so we have, sentiment analysis. 00:05:25.464 --> 00:05:35.985 So we actually read the ticket and the language that people use in tickets tells us whether there's a positive, negative or, neutral sentiment on the ticket. 00:05:36.394 --> 00:05:39.125 So we know before going into it whether people are happy or not. 00:05:39.254 --> 00:05:48.474 And so, it's kind of a interesting way to think about it because CSAT, when I was in the MSP world, was arguably the most important thing. 00:05:48.474 --> 00:05:52.485 If we have angry customers, we're already starting off on the wrong foot, 00:05:53.084 --> 00:05:55.764 CSAT is an after-the-fact measurement as well. 00:05:56.344 --> 00:06:04.164 So if sentiment measures before you go into the ticket, what are they measuring the sentiment of the person who put the ticket in? 00:06:04.164 --> 00:06:06.534 Are they measuring what happened on the last ticket? 00:06:06.584 --> 00:06:08.344 Where does sentiment come into the play? 00:06:08.394 --> 00:06:14.435 it measures the language used on, opening and on closing of the ticket. 00:06:14.485 --> 00:06:24.464 Ticket gets opened up, then you're like,"Hey, I have this issue again." Pretty bland statement except for the"again" concept. 00:06:24.495 --> 00:06:31.534 For us, that would be a negative sentiment'cause they're obviously frustrated because they've opened this thing yet again, right? 00:06:32.144 --> 00:06:38.884 and then on completion, when the ticket is done, when people respond back to the ticket, we also take the sentiment there. 00:06:39.485 --> 00:06:44.865 And, we analyze the life cycle of what that ticket did, right? 00:06:44.865 --> 00:06:48.944 So if it came in as"Oh, I opened this ticket again," that's a negative sentiment. 00:06:48.944 --> 00:06:57.764 And if it was like,"Hey, great, thanks so much for closing this ticket for us," it might go up to a positive or a neutral sentiment. 00:06:57.764 --> 00:07:02.545 And so you see it move inside of our dashboard from negative, up, right? 00:07:02.605 --> 00:07:03.574 it's not perfect. 00:07:03.625 --> 00:07:08.534 Some people use language that they don't intend to, that makes things a little bit more harsh or a little bit more soft. 00:07:08.605 --> 00:07:13.254 Guardrails is the key for us, making sure that the AI doesn't, go off and do its own thing. 00:07:13.555 --> 00:07:25.634 Just read something where one of the fast food chains somewhere in North America is rethinking their AI drive-through strategy after a customer ordered 18,000 waters. 00:07:27.375 --> 00:07:28.394 Ooh. 00:07:29.584 --> 00:07:29.884 Wow. 00:07:29.944 --> 00:07:34.798 So there are definitely some considerations before we go all in on AI. 00:07:34.798 --> 00:07:37.777 But I'm surprised that the market isn't adopting it faster. 00:07:37.837 --> 00:07:42.468 We did a study using Instinct about AI adoption in the MSP space. 00:07:42.517 --> 00:07:48.127 Only 2% of those companies were talking about AI in any organized, meaningful way. 00:07:48.337

--> 00:07:53.737 Meaning there wasn't a single blog post about AI that wasn't syndicated content. 00:07:53.737 --> 00:08:01.487 Not about how we provide AI or how, city name law firms using AI to whatever. 00:08:01.497 --> 00:08:09.218 It was, generic, 1,000 MSPs posted the same blog post about AI and a specific topic. 00:08:09.767 --> 00:08:11.747 What do you think is stopping people from making that hop? 00:08:12.387 --> 00:08:13.817 It's, the unknown. 00:08:13.927 --> 00:08:24.747 Most MSP founders are engineering-minded people and so they're very good at understanding, the ones and zeros. 00:08:24.837 --> 00:08:27.077 those things are very defined. 00:08:27.127 --> 00:08:44.447 I remember back when cloud was just becoming a thing, most organizations weren't adopting it because, like, how could it be better than having an on-prem domain controller and on-prem servers and on-prem workstations and, there's no way that using the cloud is a better option. 00:08:45.197 --> 00:08:53.998 Fast-forward to today, a decade later, and everybody's using cloud solutions because they are more effective and they are more efficient. 00:08:54.018 --> 00:09:01.057 And for MSPs, they are more valuable because it's a lot less work and maintenance Same thing happened in security. 00:09:01.538 --> 00:09:08.598 I think the same thing will happen in AI, but, most MSPs are slow adopters and some of that is for very good reasons. 00:09:09.158 --> 00:09:12.258 some people just wanna see others try it and succeed. 00:09:12.577 --> 00:09:14.908 that adoption will take some time. 00:09:15.628 --> 00:09:28.138 but it does surprise me a little bit that there's not more education or more content in the marketplace about differentiation of, say, AI and, automation. 00:09:28.168 --> 00:09:39.847 That's one of the things that I talk about almost every day with people,'cause there's lots of automation tools in place, but they're not doing the things and they're not adding the value to organizations that AI would. 00:09:39.957 --> 00:09:48.268 And so there's this, lack of information in the market to help people make informed decisions, and I think that slows down adoption. 00:09:48.317 --> 00:09:50.138 AI isn't the same as automating. 00:09:50.138 --> 00:09:51.717 How do you explain the difference? 00:09:53.378 --> 00:09:59.957 so there's two different concepts when you're talking about sales around AI in the MSP space, right? 00:09:59.967 --> 00:10:05.727 There's, I'm helping you deploy AI solutions, right? 00:10:05.737 --> 00:10:15.048 And so I need to understand the pain and the value that, you're trying to create or the problem that you're trying to solve leveraging AI or automation, right? 00:10:15.048 --> 00:10:16.518 I gotta understand what you're trying to do. 00:10:17.258 --> 00:10:24.543 That is one piece of it, and that's us just being able to help with- Chat GPT or Claude deployments, right? 00:10:24.543 --> 00:10:27.443 How are you gonna use it, and what are the guardrails, and, how do we set it up? 00:10:28.053 --> 00:10:45.114 but then there's the other piece of AI, which is more MSP consumption, delivering services to end customers, and I think for the most part, most MSPs don't want to be transparent about using AI solutions. 00:10:45.124 --> 00:10:47.354 Most of it's white-labeled in some form or fashion. 00:10:47.403 --> 00:10:49.994 they're not actually selling that they're using AI. 00:10:50.024 --> 00:10:51.303 They're selling the outcome. 00:10:52.234 --> 00:11:08.703 "Hey, we can solve your tickets 40, 50, 70% faster, than the MSP down the street." Now, they know on the back end the reason they can do that is they're leveraging AI to solve those tickets faster and more efficiently. 00:11:09.254 --> 00:11:17.183 so they're using that marketing capability to go and win business by outpacing and out-delivering their competitors in the space. 00:11:17.913 --> 00:11:31.443 but I don't think anybody's really going out there saying,"We're using and delivering AI to our customers," because again, there's still this mindset that AI is bad, right? 00:11:31.783 --> 00:11:33.244 AI's taking everybody's jobs. 00:11:33.274 --> 00:11:35.313 AI's doing all of these bad things, you know? 00:11:35.423 --> 00:11:37.173 and there's some bad AI stories out there. 00:11:37.173 --> 00:11:46.754 Like, there was a story not too long ago with, an AI solution that, completely deleted a server and then deleted the backup to cover its tracks, right? 00:11:46.803 --> 00:11:48.933 that's not a good look for AI. 00:11:48.963 --> 00:11:57.724 A lot of organizations don't have guardrails to solve for that, we have humans in the loop in the back end to help make sure AI does its job. 00:11:58.264 --> 00:12:07.543 And so there's never a scenario where the AI's just, going out and doing everything on its own without having humans check in and make sure that it's staying in line. 00:12:07.604 --> 00:12:10.713 So there's some of that still out there in the market. 00:12:11.693 --> 00:12:14.374 Is it a fear of not understanding AI? 00:12:14.894 --> 00:12:21.963 Is it a fear of having to change price points if people now know that you're using AI instead of human labor? 00:12:22.533 --> 00:12:24.413 Where do you think the fears are coming from? 00:12:26.138 --> 00:12:27.687 I think both of those are accurate. 00:12:27.758 --> 00:12:48.298 one of the things that I've always found interesting is that, no matter the size of the MSP, generally the value proposition that they're selling is that they're able to solve a problem, and the, kinda bring it home value prop is that you have people right down the street that you can pick up the phone and call, and they can come and help. 00:12:48.927 --> 00:12:55.418 That local touch and feel for most MSPs is a huge value creation concept. 00:12:56.057 --> 00:13:10.018 so if you are saying,"Hey, we're using AI to solve some of these issues," there's a disconnect between that value proposition of

somebody down the street, and then AI's actually delivering the solution. 00:13:10.107 --> 00:13:33.268 And so I think the next iteration of sales techniques and tactics is going to be how do I create this value proposition in more of a hybrid environment where I'm selling the efficiencies and effectiveness of AI to maybe reduce your price point or to make you more efficient and get you back up and running faster. 00:13:34.107 --> 00:13:40.258 But also, I have people down the street that can come out and help you if we run into that situation, right? 00:13:40.967 --> 00:13:43.487 A full messaging shift almost. 00:13:43.977 --> 00:13:45.018 It feels like it. 00:13:45.097 --> 00:13:56.697 it may not be a completely full pivot, but I think there's gonna be a pretty substantial messaging change over time as AI becomes more prominent, and it's obvious that MSPs are needing to adapt to it. 00:13:57.307 --> 00:14:16.268 the reality is that, most platform organizations or, the larger MSPs with PE backing are already starting to look at AI because, the fundamental truth of business is that you're trying to protect EBITDA margins, and you need to deliver on profitability and, client satisfaction. 00:14:16.317 --> 00:14:23.337 there's already a push from a lot of the private equity groups that are saying like,"Hey, AI is a thing. 00:14:23.597 --> 00:14:25.737 I wanna see how it works in this space. 00:14:26.357 --> 00:14:36.717 Maybe it doesn't, maybe it does, but, we need to start looking at it and start to figure out adoption of it to make things more effective and more efficient." So you're already starting to see that. 00:14:36.998 --> 00:14:56.967 once they are really good at it, it's gonna end up driving price point down on more regional and local players because the larger MSPs are going to be able to deliver tier one services at substantially less, dollars than, the local and the regional providers doing it with humans. 00:14:59.035 --> 00:15:10.826 You know, there's already AI solutions that I was just mentioning it before, when I call into the hotel and I got an AI chatbot, and they were talking to me like a human. 00:15:11.426 --> 00:15:22.245 we're developing that internally as well, so, that's something that we will be offering, very shortly, where people can call in and talk to what they perceive as a human. 00:15:22.436 --> 00:15:26.355 or we can just have chat, you know, typing messages back and forth. 00:15:26.405 --> 00:15:28.306 most people don't wanna talk on the phone anymore. 00:15:29.015 --> 00:15:38.525 the generations that are coming into the market that are starting to lead the charge in technology and adoption, they don't wanna communicate via phone. 00:15:38.535 --> 00:15:39.905 They wanna communicate via chat. 00:15:39.956 --> 00:15:44.336 And so in that world, AI is pretty dominant. 00:15:44.336 --> 00:15:47.836 It can do a lot of things in chat, functions. 00:15:47.936 --> 00:15:51.926 Just like everything else that comes into

the MSP market, there's this big yo-yo effect, right? 00:15:51.926 --> 00:15:55.235 So there's, low adoption, very low engagement. 00:15:55.316 --> 00:15:59.725 Then there's, heavy adoption, and then, the pendulum swings all the way to the other side. 00:16:00.466 --> 00:16:05.966 Then everybody realizes we pivoted probably too hard, and we need to come back to the middle somewhere. 00:16:06.015 --> 00:16:08.936 we haven't gotten to the heavy adoption yet. 00:16:08.946 --> 00:16:18.755 There will be a pendulum swing where it's gonna be too much, and then, pretty soon, I would say probably next year, you're gonna see some of that, evening out. 00:16:18.926 --> 00:16:29.795 I've also seen a lot of articles and a lot of people talking about 2027 being the year of people doing homegrown AI, starting to catch up with them. 00:16:29.855 --> 00:16:40.466 there are organizations out there that are building their own, by coding through Claude or through something else and building it, but maintaining and managing it is actually really difficult. 00:16:40.515 --> 00:16:50.706 And so the ones that started last year or even this year, next year is when the maintenance and the management really starts to catch up with them. 00:16:50.706 --> 00:17:04.086 And so the idea is that those people that started that, unless they have the money for dedicated resources to manage it and maintain it, are going to be on the market for new AI solutions that are actually managed by someone else. 00:17:04.936 --> 00:17:16.470 So I also think we're gonna see a dynamic shift there too, where- There's not a lot of organizations out there that can afford to manage and maintain a homegrown AI solution. 00:17:17.059 --> 00:17:19.190 It's just like PSAs back in the day, right? 00:17:19.870 --> 00:17:22.130 You have a couple of, major PSA players. 00:17:22.450 --> 00:17:24.140 Some people tried to build their own. 00:17:24.830 --> 00:17:33.750 Once they realized that they can't manage and maintain it effectively, they ended up buying from, a Connectwise or a Kaseya or a Halo or, one of the other big guys. 00:17:35.250 --> 00:17:36.620 So we're gonna be in the same spot, I think. 00:17:38.150 --> 00:17:45.860 How do people create a, risk mitigation strategy for the things that they're building right now for 2027 and beyond? 00:17:46.750 --> 00:17:50.019 Yeah, that one is a tricky one, too. 00:17:50.100 --> 00:17:52.569 I'll take this back to the education and knowledge piece. 00:17:53.600 --> 00:18:03.339 I think a lot of people don't know what they don't know right now, so they're building, and they don't know what risks that they need to be looking for, and there's not enough track record to figure out where the pitfalls are before they build it. 00:18:04.339 --> 00:18:10.420 So I think people are, running headfirst into some very real problems that they're gonna have. 00:18:10.960 --> 00:18:22.829 For us, we built a product with the intention of going to market with AI and understanding the AI economics of token costs and things of that nature. 00:18:22.829 --> 00:18:34.539 And so we've put in lots of different things to make sure that even if token costs go up with one or two of these, larger providers, we have flexibility on making sure that we can pivot and adapt to that, right? 00:18:34.599 --> 00:18:37.599 We already knew that was a pitfall, and so we built for that. 00:18:38.569 --> 00:18:43.220 I just think that right now a lot of organizations aren't building and taking into consideration that. 00:18:43.289 --> 00:18:52.410 Or the pricing model for the AI product is based on token costs, and so as token costs go up, so will their price point. 00:18:52.700 --> 00:19:00.970 Whereas, groups like us that are built around that mindset, we're able to manage our price point effectively in a pretty linear fashion. 00:19:01.599 --> 00:19:11.500 obviously inflation and things of that nature come up, and you end up having to raise prices slightly, but, you're not doubling or tripling or quadrupling your price because now token costs are so much more. 00:19:12.430 --> 00:19:15.059 and that's just one aspect of those pitfalls, right? 00:19:15.109 --> 00:19:18.099 I mean, to your point, there's a lot of them out there. 00:19:18.589 --> 00:19:23.440 I don't think they've been uncovered yet for a lot of these organizations to see. 00:19:24.009 --> 00:19:26.720 there's groups like us that have been doing it for a while. 00:19:26.720 --> 00:19:27.180 We've seen it. 00:19:27.190 --> 00:19:31.440 We've built some stuff internally in enterprises, which is how this organization started. 00:19:32.059 --> 00:19:34.519 we've seen a lot of those pitfalls, and so we've built for them. 00:19:35.740 --> 00:19:41.009 if you're an MSP, you're gonna start talking about AI in some way. 00:19:41.109 --> 00:19:44.029 What kind of content should MSPs be creating? 00:19:44.500 --> 00:19:53.220 right, wrong, or indifferent, the way that I've always operated in the sales world is, Identifying problem and pain points, and then trying to solve those pain points. 00:19:53.750 --> 00:19:58.400 in some sales organizations you see people that try to, square peg, round hole, right? 00:19:58.470 --> 00:20:01.190 And that doesn't necessarily work. 00:20:01.220 --> 00:20:06.359 My two cents is always going to be focusing on problems, like real-world problems that people are running into. 00:20:06.359 --> 00:20:13.410 So the things they should be talking about are, you know,"Hey, my customers are a law firm," for example. 00:20:13.809 --> 00:20:16.190 And, I just saw this the other day. 00:20:16.529 --> 00:20:19.299 law firms are predominantly like professional services orgs, right? 00:20:19.319 --> 00:20:20.329 They bill by time. 00:20:21.000 --> 00:20:35.599 Well, they use AI internally to make their time more efficient, but if they're billing for time and AI is making them more efficient, they're actually losing money because they're billing less hours

for that time. 00:20:36.299 --> 00:20:57.349 So a blog post that I just recently saw was an MSP talking about how, they helped a law firm change the way that they were billing to make sure that they were still, billing the right amount from a revenue standpoint, still getting the right amount of dollars in the organization, still capturing the time savings from the AI solutions. 00:20:57.779 --> 00:21:00.000 So it's not always just about the technology, right? 00:21:00.000 --> 00:21:06.559 It's like, how do you use the technology, and how does that technology become relevant to your current business model? 00:21:06.559 --> 00:21:08.549 And does your business model need to change? 00:21:09.059 --> 00:21:31.329 And if so, as an MSP, we see a lot of different types of customers and a lot of different things that are happening, so we can bring some of those knowledge points and education points to bear and say,"Hey, not only can we help you from the technology piece, but we can help you as a business advisor," because functionally at the end of the day, that's what we do in sales, is advise people. 00:21:31.900 --> 00:21:35.609 it's not always just pushing the product and solving, the revenue goal for us. 00:21:35.609 --> 00:21:37.420 It's like, how do I solve a problem? 00:21:37.420 --> 00:21:54.200 And so to me, that's what the marketing and the, education in the space should be, is how have we, you know, as a person that delivers the service, how have we seen this change the market, and how have we seen AI impact it, and what can we do to help in that scenario? 00:21:54.200 --> 00:21:55.759 Or what have we done to help? 00:21:56.160 --> 00:22:06.720 I think that messaging goes a lot farther than, hand waving and saying like,"Hey, AI can do all of these great grandiose things," because sometimes it's not relevant, right? 00:22:08.723 --> 00:22:10.963 I think that's a great way to, wind down for the day. 00:22:11.423 --> 00:22:15.574 thanks for coming on Win and sharing your, thoughts with us today. 00:22:15.584 --> 00:22:18.824 where can people reach you and learn more about Fixify? 00:22:19.794 --> 00:22:20.054 Yeah. 00:22:20.054 --> 00:22:22.693 Actually, you can just go to the website, fixify.com. 00:22:22.723 --> 00:22:25.433 There's, ways to schedule demos with me. 00:22:25.433 --> 00:22:28.874 There's ways to grab, educational content and information. 00:22:29.433 --> 00:22:37.104 we actually have an AI chat bot that you can talk to that has real questions, and answers there, so, you can get a lot of information from her as well. 00:22:37.114 --> 00:22:40.394 If you go to the website, Fiona will pop up in the bottom right-hand corner. 00:22:40.493 --> 00:22:43.483 and you can have an actual real conversation with her. 00:22:43.773 --> 00:22:52.874 Tell her, what size organization you are, who you serve, what you do, and she can actually have real conversation about whether we're a good fit and how we would be able to help. 00:22:57.109 --> 00:22:58.180 Have a good afternoon, Matt 00:22:58.549 --> 00:22:58.809 Yep. 00:22:58.970 --> 00:22:59.430 Thanks