The MSP 3.0 Shift Is Here

Speaker A: We need to evolve. Our market has evolved, whether we like it or not. I miss writing code, but I can guarantee you that other than very special circumstances, I probably will not write another line of code the rest of my life. Whether we want to admit it or not, the world has changed. And it's our choice whether we choose to change along with it or to try to cling to the old ways.

Speaker B: Hey guys. Damian Stephens, host of MSP Mindset, founder and CEO of Servocity. Today I am blessed to interview Jerry Miller, CEO at Cloudticity. Now he's known for being one of the earliest MSPs to cloud and one of the deepest adopters building out systems around it. Now he has done the same with agentic AI. And the can't miss part for me for this conversation was how do you go from chatting with AI to using it to automate a little bit, to automating just about everything you can imagine? Which has introduced joy for his team to the stage they're at now where they're imagining things that were just impossible before. Things like proactive detection, before Churn even becomes a real risk. If you'd like that and you'd like to understand his journey, his deep dive into agentic AI. Don't miss out on our conversation today. Jerry, I'm excited to dig in today. Thank you for being on a Ms. P mindset.

Speaker A: My pleasure. Thanks for having me.

Speaker B: So we share a lot of, I think common beliefs, but one of the things I feel like most people are missing is how deep you've gone into AI, and specifically agentic AI. Um, let's dig in there. Like, tell me what that has meant to you and your company, your team.

Speaker A: Yeah. So I guess the headline there is, is that we in general as a society are inundated with information from many, many sources.

Speaker B: Yes.

Speaker A: And at a, ah, deeply rooted in technology company like Cloudticity that, you know, even though we're in our fifth year, we started as cloud native, we started as remote. We use all uh, sorts of tools in our everyday life. So our inundation with information from a wide variety of disparate sources has been a blessing in that we're able to build a remote company very quickly and get to revenue very quickly and start producing positive impacts for our customers. Uh, but internally it means that we have to track a lot of pieces of information from a lot of sources. And uh, until we really got facile with AI, that was primarily a manual effort. I remember to go check this and go log this time over here. And this other system. Oh, I forgot about it. Um, so with the AI tools that we have at our fingertips today, really for the first time in our company's history, for the past six months, I would say we've been able to integrate all of our Data sources from 20 plus SaaS, applications, um, all the way down to recordings of conversations. So voice, email, Slack communication, time tracking systems, financial systems, HR systems, employee satisfaction, every single aspect, every signal of our business is accessible at our fingertips now. And we can begin to connect dots, uh, and draw inferences that were actually absolutely impossible six months ago. Which means we can now automate detection and response of all sorts of signals, some of which we don't even know about. And that ability to build those systems and automate them with agentic AI has been absolutely game changing in our ability to run an efficient business as well as service the clients that rely on us.

Speaker B: I get the pleasure to be in build sessions with MSPs just like you building real things with AI. And the number one friction point, the number one blocker is I can't connect to all the systems. In fact, my vendor makes an MCP server or a connector or whatever they call it, yet it's read only another one tried it and it limited them to 100 records. How are you going to search for all the tickets when you can only get to 100? If you're tired of vendor connectors and MCP servers holding your business back, check out the the MSP Skills repo. The link is below. It's got over 50 different MCP servers, skills and connectors that you can use to connect your entire stack to the agentic AI of your choice. Now why do you care? The reason is you can now go through and say find all the unused licenses that I'm paying for that I shouldn't be. Would you like it to go through and prep across all your different tools for that qbr, that thing that took you hours to prepare for for one client is now one sentence. So if you want to make sure you're getting the value out of AI, you've got to connect it to your real world business systems. And there is no better place than the MSP Skills Repo. Check out all the link below. And if you have questions or if you just want to join, learn more, come to the one of the build sessions. You'll see us putting it into action. So you said a ton there. There's a ton. A ton to unpack. I love that you started with right. We, we've kind of Long been in the information age and somewhere along the lines somebody forgot to tell me. We're in the information overload age. Yeah. And uh, it's kind of like the, the frog in the, uh, boiling pot. And yes, finally I realized, um, so help me. Like, I know you do a lot personally with AI.

Speaker A: Yeah.

Speaker B: Which came first? Was it the personal journey or the team journey or some other way?

Speaker A: Uh, it was the personal journey. Not just me, but a few of our vanguard players. Everybody at Cloud City is a deep technologist. And so we're all experimenting and playing, including me. Um, and so I had my experiments, other people had their experiments. Um, about, uh, maybe eight months ago or so we realized that we should probably collaborate loosely. And so we want to provide the ability for people to collaborate without putting rules in place. You have to do this, you have to do that. Right. So we want that collegial, uh, atmosphere that promotes creativity but allows us to work together and begin to coalesce around common standards not through edict, but sort of through natural process. It's about eight months or so we started the AI, the Cloudticity AI Venture Lab, which is a playground that allows people access to a wide variety of AI tools, a set of internal channels that we can communicate. Um, every other Friday we all meet for an hour and kind of do a show and tell and share best practices. And that has really sort of driven a collaborative effort. And we've begin to, as, as an internal community, coalesce around, uh, I don't want to say standards in a way that we uphold, but best practices, things that work well, coupled with deep investments where, you know, everybody at the company has a cloud teamwork or team subscription. And so we're big users of cowork, um, we're big users of cloud code, but we're not limited to that. We also have API access, um, to any backend AI system. And so we've got API access to OpenAI models, open weight models, you know, um, we're a big Google Apps shop, so we've got Gemini access. So, um, providing access to a wide variety of AI tools, both programmatic as well as you, uh, know, just sort of application oriented, coupled with this, uh, organic, um, effort toward building an internal community to continuously update our knowledge and support each other has resulted in some widespread internal AI adoption that we're seeing tremendous, uh, internal and external results over.

Speaker B: So I guess this was opt in. You created this and anybody who wanted could opt in. Um, do you have any lessons learned or any kind of tips you would share for Encouraging people to do that and experiment, uh, especially when some people are natural into that. But what I have found interesting is this seems different as a technology and the people that are always into the next thing sometimes aren't. And sometimes there's fear and sometimes there's excitement and curiosity and all of those things.

Speaker A: Yep, yeah. And we see like I personally experience all of the above, as I imagine most people do, uh, we have an intrinsic advantage in two ways. One is we're a tech first company. Right. We are at heart all uh, engineers, even those of us that are in non technical roles. And so we live and breathe bleeding edge technology. And we cut our teeth starting 15 years ago at uh, building the healthcare solutions that others wouldn't. And we've always blazed trails and that's been a deep part of our culture from day one. So AI and taking on something that's a little bit scary, a little bit uncharted, a little bit new is sort of natural to us. Um, and the other thing is that because we operate in such everything is new environment, we're blazing new trails, we have to be tenacious and courageous. Um, we tend to attract and hire intensely curious people, people who just, they'd love to go to sleep, but there's something new on the horizon they've just got to go learn about. Um, and so those two attributes of what it takes to be a successful cloud, titian, which is how we refer to ourselves, um, have been natural, um, buoys as it comes to uh, incorporating this completely disruptive technology, uh, that has been thrust upon the world.

Speaker B: Mhm. Speaking of completely disruptive, how did you go from experimenting and a bunch of people getting excited, which I love. But you've got healthcare clients, clearly you have governance and regulations and all, you have all these things that, that's where some more fear comes in of like, well, I can, I can get this account and tinker with this. But you know, there's a difference between that and using that, you know, either internally or for clients or in any production manner. So how are you able to do that? And like you mentioned earlier, like this is innovations, not quarterly, it's continuous. I love that. And so how are you able to do that kind of at the speed that might be needed in kind of an AI era? Because I think we're moving past the um, once every year or three, reevaluate vendors.

Speaker A: Yes. Yeah, it's pretty continuous. It's a good question. Um, we think about, uh, concentric circles of trust. Um, and so for me to take Claude Cowork and connect it to my personal Slack account in my calendar is pretty low risk, right? The highest risk is it might decide to go send a bad slack message to people right on my behalf. So there's very, there's little to no client risk there. As we move in inward in those concentric circles, all of a sudden now we're at a client's AWS or Azure or Google cloud account in which their protected health information resides. We don't touch those things. And so you know, I said we have, we try to maintain a creative rule free environment and at the outer edge of those concentric circles we have guidelines and best practices. At the inner circles we have extremely rigid hard rules and those are vetted by external organizations. We've been high trust R2 certified coming on almost a decade, uh, SOC 2 type 2. Um, so we have external teams validate that we're not doing dangerous things. So our AI tools never touch client accounts. We don't manage our clients cloud accounts with our AI, we connect our, our business systems, our email systems, our accounting systems, but never manage clients. The only teams at cloudticity that are allowed to introduce AI into client facing environments, um, uh, it's one team, it's our platform team that is responsible for development and management of our Cloudticity Oxygen platform which manages our clients accounts. And even then they're extremely judicious about how we apply AI. So it's a read only. Uh, we don't write back into accounts, we don't look at data, we only look at metadata. So we're not looking at protected health information, we're looking at is the account configured in accordance with high trust best practices. Um, and it's always human in the loop, it's always read only and it's never accessing actual data. And we vet that. Organizations like hitrust, where we're deeply involved, have introduced extensions to their security frameworks. There are almost 50 new controls from Hitrust around appropriate AI management actually was on the committee that developed those rules. And so we remain deeply involved not only in the vanguard of how to use AI at the cutting edge of technology, but how to use it safely and responsibly. Understanding the different areas of risk around, you know, maybe it messes up your email, but that's a whole lot different than it releases 100 million patient records.

Speaker B: Right? I love that because I feel like some people fall into the we're highly regulated or we only work with highly regulated, so therefore we can do nothing.

Speaker A: Right.

Speaker B: Um, and I think that is too binary of a way to look at it.

Speaker A: It is by far. And the investments that we've made, both monetarily, but much more importantly, uh, the investments of time that our team, nights and weekends, just because they're so excited about this, has yielded tremendous productivity gains at zero additional risk to our clients or the patients that they serve.

Speaker B: So tell me about some of those gains. Um, I don't know if you want to start about a project. I know you have had huge gains and what you're able to do per person, all sorts of other things. I don't know the best way to even start to peel the onion, but you're no longer chatting with AI. You're doing a lot more.

Speaker A: Yeah, yeah, yeah, yeah. It's all about those agentic loops. At this point, much of what we do with AI is automated, um, either scheduled or triggered tasks, uh, agentic loops, um, automating away a lot of the toil that takes away from the things that humans can uniquely bring to the table. So there isn't a world where we're looking to replace humans. But what we're trying to do is to remove those redundant tasks that take a lot of time and effort, but really don't add a tremendous amount of value. And for us, the way that we think about value is have we helped our customers serve their patients better? You may remember from our last conversation at Cloudicity, our vision is to help every human on earth get healthier through the work that we do. And we're not clinicians, but we enable clinicians to do their jobs better. Uh, if I'm filling out some report that is not helping our clients serve their patients better, it's toil. So those are the kind of things that we look to automate away with with AI and agentic AI to remove the toil that prevents us from doing the things that today only humans can do, which is bring creativity to the table. Think of unique solutions, think of opportunities that we can bring to our customers to make their healthcare businesses more efficient and more effective for the patients that they serve. That level of imagination cannot be replaced by AI, but the toil that stands in the way of that emission of that imagination, um, can. And that's how we think about leveraging AI to drive efficiencies in our human performance.

Speaker B: Can you give me some examples of how you're driving that? Because I think some people, if you haven't gotten your first kind of win, it's hard to imagine.

Speaker A: Yeah, we look at things that are time consuming, that don't need to be time consuming. So we do, for some of our larger clients, we do monthly business reviews. For smaller clients, we do quarterly business reviews. But these are not sort of check the box activities. We spend a tremendous amount of time ensuring that that review is of significant value to our customers, who are really devoting their time. And many executives show up. We want to make sure that that's a valuable exercise. And it used to take somewhere between 18 and 25 hours to prepare an MBR or a QBR. And, you know, if you have, if you cover between six or 10 customers, that's a lot of time. Um, and much of it is like your entire month. That can be an entire month. And that's just preparation of a meeting, not doing the actual work. And so much of the research that goes into it, we have fully automated away. And so, um, our systems know our customer schedule, whether for an MBR or a qbr, and they're able to pull data from a wide variety of sources. Every phone call or zoom call we've ever had with a customer, every email, every slack message, uh, their, uh, security and compliance posture. Over time, the JIRA tickets, they may have opened our performance on satisfying those tickets. Every aspect of how we have handled that customer, what's going on with them, coupled with external research. Did they enter a strategic acquisition? Uh, were there press releases either good or bad? Um, did they do a, you know, an SEC filing? Um, did they raise money? Like, all of these things take time. So we have fully automated the preparation of those business reviews now through AI and all the way down to scheduling cadence of communication. Because it's not just preparing a deck, it's getting that deck out ahead of time. By the time we get to the meeting, we're not going through a slide deck. We're talking about what that means and what are the next steps. So management of that entire thing, from prep to delivery, is now fully automated through agentic AI loops. And that saves dozens to hundreds of hours in a month for, uh, all of our cloud value architects, freeing them to do the creative things to really process. Okay, what does this mean for my customer? And how can we make sure that next month's review is even more accretive to their core mission? Um, and so freeing that time and allowing us to repurpose that time to more value add activities has been a tremendous value add to our team as well as to our customers and the patients that they serve.

Speaker B: Yeah, um, speaking of the team, what's been the reaction? Because again, I see, of course, there's earlier adopters, maybe there's folks in the middle, maybe there's laggards. But is it generally seen as this is great, this is reducing toil, or did you have the initial fear of um, replacement or um, taking some of the things I enjoy away or other, other issues we, we've had.

Speaker A: You know, again, as a whole, if you're a cloud titian, you're excited about technology, you want to play with the shiny toys, um, and you have a deep sense of curiosity about this new stuff. So while I talk to many CIOs who are investing heavily in AI and having trouble driving universal adoption, we're blessed with not having that problem. Um, and that being said, there are those that have adopted it more quickly and those that have not. There are those for whom the adoption was really easy, almost natural, and those for whom it's a culture shift. Um, so we haven't had a ton of pushback, we've just had different pace of adoption across the team. And that's okay, that's to be expected. Um, one interesting side effect is that we've skinnied the team down quite a bit. Not because we've let anybody go know you hear about these massive layoffs to be replaced by AI, uh, that's not the cloudticity way at all. But you know, through natural attrition, we've realized that there are certain roles that we just haven't had to replace. And so we're probably down in the past 18 months, probably down about a third in headcount. Um, not through layoffs or anything along those lines, but just the efficiencies that we've experienced through Uviqua's adoption of AI have resulted in a team that's so efficient that we don't have to hire every single time we lose somebody. Um, and we're still, we have open roles and we're still hiring, so we're still growing, but we're not as desperate to fill chairs as we have been in the past. Um, in the pre agentic AI days.

Speaker B: Yeah. So for any of you guys wondering, listening, like if you think about that, you're telling me with a third less staff, you're still able to accomplish the same mission. And it sounds like, not just accomplish in certain areas, qbr, it sounds like you're doing it faster or better.

Speaker A: I think we are, I think we're with a smaller team, we're executing better, faster, cheaper. Um, we know our customers better. Um, in addition to the qbrs or mbrs, we have added AI based continuous research into our internal tooling. So every time one of our folks logs into our systems, they're seeing proactive updates about what's happening with their customers. We're looking at signals of potential customer dissatisfaction long before it becomes a churn. Eventually, um, we're building automated customer next steps. So everybody wakes up to an AI generated dashboard. You know, here are the top three things that are most important for you today, which in many cases are counterintuitive. Um, you just wouldn't connect certain dots that we're able to now. And so I would argue that we're probably even more efficient and even more customer focused than we've ever been, um, despite sort of a natural shrinking of the staff just through natural attrition. And we'll continue to hire and continue to grow, um, but at a much more measured pace than we've ever been able to in the past.

Speaker B: Mhm. You said something interesting because a lot of people I think are what I would call chatting with AI. They don't have experience yet in agent like AI, but you. But I think then a lot of people jump to just automating um, workflows, which is a great place to start. You talked about that with QBRs, but then you mentioned like signals for what to do signals before churn becomes an event. And those are probably not workflows that really even existed that were probably not possible before.

Speaker A: Yeah, yeah, yeah, it's a really good point. You know you, you start with chatting. The next step is generally having a conversation with your data. So tools like Notebook LM are great to load up a lot of your stuff and you can actually have a conversation with your data. Including now with tools like fathom that we are big fans of. Your data might include phone calls or conversations, um, and then using the ability to connect those dots. The next logical step is to take your existing workflows and begin to automate them. Which you can do now by describing things in markdown plain English instead of writing code. So the automation of tasks becomes much more accessible. And then the next logical step from there is to begin automating things that you didn't know existed. And so looking for signal in what used to be a pile of noise becomes a really powerful application. And not knowing the rules, but having AI begin to derive the rules by doing regression analysis and that sort of thing to begin to glean insight in non heuristic ways. And that's the extension beyond automating known workflows is to begin to automate unknown workflows. Um, through automated signal detection. And that's where agentic loops become exceptionally powerful.

Speaker B: M. Mhm. You're very cutting edge in all of this and certainly with agentic loops, that's the uh, the cutting edge of the month. Um, topic. Uh, tell me more about that. And what I mean is like what does that mean in terms of what do I do with that from a business perspective? And just um, what are you guys thinking about? Like where do you go to apply that? You know there's this loop and it can continue to improve things is the generic. But I'm going to let you answer this. But, but then like what do I do with this newfound power? Uh, I think is often what I think.

Speaker A: Well we, you know, it's as a techie, right, Like I'm up at 3:30, 4:00 clock on Saturday mornings because I get five hours of playtime before the kids are up, right. So yeah, um, and I relish that time. So. So I love to play with technology. But when we apply this to business, we always start with a question, right? We might start with um, how, uh, you know, our customer churn, we love our churn, make customers stay with us for a long time. But how can we cut that even further? How can we identify a morale internal problem before it becomes a problem? So we ask a lot of what if, right? What if what? What if we could improve our margin by X? What if we could reduce customer churn even further by Y? What if we could generate absolute daily delight amongst the cloud titians so that they just love to be here, which is a dream of mine, is to build a place where people love to be. So a lot of these what ifs turn into okay, how are we going to do that? Because these are things that were close to impossible before. They would require heroic efforts and um, and heroic budgets. And heroic budgets, right? And, and many of these things are lagging indicators like, you know, you don't like. Generally your first indicator of customer churn is when you get the letter, the termination letter, right? And you know, so, so this technology has allowed us to turn lagging indicators into leading indicators by reimagining pieces of the business by asking what if? And then from there the. That's where we begin to apply the technology because we don't have to solve that problem, as it turns out. So the thing I love about agentic loops is you're feeding information about your business into the agents and you have a set of agents collaborating who can begin to develop hypotheses that other Agents loop and begin to test those hypotheses and that can run on its own. And so you've got smart agents that generate smart hypotheses and then other agents that disprove those hypotheses until they come upon one that they can't disprove.

Speaker B: Mhm.

Speaker A: And then that we promote into real production and then we have other loops that track accuracy. So what did it get right? What didn't it get right? So the concept of a loop with agentic, uh, AI is really about hypothesis generation and testing and then continuous improvement based on real world results. And that technical aspect applied to real business problems is where magic is starting to really happen.

Speaker B: Yeah. What are you most excited about now with all this new um, capability? What I frame as, uh, the question I keep asking myself is what was impossible before?

Speaker A: Yeah.

Speaker B: That now may not be.

Speaker A: Yeah. Uh, one of the most exciting and most frightening aspects of AI is the accelerating pace of innovation. Now Claude code writes quad code.

Speaker B: Mhm.

Speaker A: Which means that the pace of releases is accelerating. The same is happening with the AI models. Most of the AI code is each model writes its successor and there's still a lot of human in the loop. But that's diminishing to the point where we've had to put manual brakes on this pace of innovation. As you look at mythos and fable. Right. Those were too powerful to be released and we didn't have bureaucratic guardrails or society guardrails. And so we are literally putting manual brakes on this pace of innovation because we don't know how to absorb that pace of innovation.

Speaker B: Yeah.

Speaker A: As we begin to evolve policy and evolve guardrail technology, uh, and evolve usage guidelines and characteristics, we will more and more be able to absorb that accelerating pace of innovation. And that really excites me because every day is going to bring something that we can do today that we couldn't do yesterday. And as an entrepreneur, as a business owner, as a passionate believer in technology, being able to having the capacity to make humans better and in our case healthier, if we can figure out sort of the responsibility and the governance that has to come along with this, this possibilities of uh, endless future keeps me up way later than I should be most nights.

Speaker B: I want to ask a different question because you are cloudticity. You guys were very early in cloud and not just early. I think there's early and um, then there's adoption. Uh, and you were both, you saw, I think that it was different. Lift and shift was not going to work, for example. Um, and so you've built an amazing business by you driving through that. But my question is, I talk to so many people and they're like, yeah, this is like cloud or cyber security or something. What's your take on that?

Speaker A: This is like cloud or cyber security,

Speaker B: meaning AI or AI and agentic AI. Like how does it, is it different? And if so, how is it different?

Speaker A: It's dramatically different from anything I've seen. And I've been in tech more than 40 years. Um, and so I was there through the birth of the Internet, late 80s, early 90s, and built businesses around that and they were successful. And we thought that was the greatest breakthrough ever. It came so fast and it changed so many things and disrupted the world so quickly that we could never imagine such a cataclysmic change. And then AI hit, um, and it's been around forever. You know, the concept of AI and the concept of neural networks has been around for decades. Um, but a sort of tidal convergence happened that made it real. So we have computing capacity, we have suddenly these GPUs that were designed for my kids to play games happen to parallel process in ways that neural networks like, um, we have cheap memory. Well, it's not so cheap anymore, but it'll get cheap again. We've got storage that's nominally free. We've ubiquitous connectivity. Um, we've got, you know, we're on the break of some significant energy breakthroughs. Um, but we do have good renewable energy capacity. And so all of these breakthroughs converged to make AI possible. And it's even been around for a while. Open. OpenAI had GPT forever, but somebody just thought, let's throw a chat interface on it. And you know, it grew to 100 million users faster than any technology ever. Ever.

Speaker B: Mhm.

Speaker A: And it's taken over the world. And now that it's writing itself, it's going to continue to take over the world. So this pace of change is unprecedented compared to any technology, not just computing technology. It's so disruptive and groundbreaking and it's going to change so many things. Like we, we have to make sure not to let AI think for us or we will lose the ability to think. So we have to have discipline to use AI judiciously. Kind of like, you know, using the calculator doesn't make you stupid, but you really still have to understand the fundamental concepts of math. And the calculator has to enable you to do more advanced math that still requires you to think. So we have to be responsible and disciplined in our use of AI. To make sure that it's an additive technology, not a replacement technology. As we can see, AI has been incredibly disruptive to the job market and will continue to do so. So we're going to have to rethink regulations and societal norms and even income tax. Uh, every aspect of our economy is based on people working and people consuming. And that equation is going to shift dramatically. It already has started. Um, and so we're going to have to devise new policy and new regulatory frameworks to accommodate this new reality. As AI begins to or continues to embed itself into the physical world, Robotics, those are becoming ubiquitous and cheap. MCP servers and skills and such that connect AI, which used to be a chat interface, to every system that we manage. My AI turns my lights on and puts my shades up and down and pretty soon it's going to drive my car. These are things that we have no societal experience nor frameworks for.

Speaker B: Mhm.

Speaker A: And we are going to have to scramble to catch up with the pace of innovation in AI to match that pace with policy updates, regulatory updates, um, fiscal, economic updates. And those are the pieces that we as a society just don't have answers for yet. Right.

Speaker B: What do you think we may be missing when it comes to what's been the managed service provider industry? Right. Uh, and some of us are still managing infrastructure and some of us are more into the cloud or cyber. There's all these different things, but obviously everything will change. Um, most things will change. So what do you think the opportunity is and what do you think is going to change?

Speaker A: That's a great question. Um, we saw sort of a coming shift in MSP managed service provider business about five years ago, pre AI, and it's been accelerated by AI. So if you think about the managed services industry, if you think about MSP 1.0 was, I believe the term back then was called your mess for less. So an MSP would kind of take over a company's IT operations and theoretically deliver it more efficiently. And IT still required armies of people. And some, you know, offshoring really became a thing during this period of time. But armies of people in data centers racking servers or answering help desk calls. Um, that morphed when the cloud came into being. Started with virtualization and then really cloud, where physical infrastructure became relegated to somebody else. And what used to require a human racking a server, for example, became an API call mhm. And the MSP 2.0 was the era of automation where you used to have to have a person put a server in a rack now you have a piece of software called an API to virtually provision a server. And the MSPs that were early on, that bandwagon did exceptionally well. And the MSPs that stuck to the old ways, um, didn't do so well. That automation, that ability to be a magician, allocate a terabyte of storage in seconds that has become somewhat commoditized. And so we foresaw, ah, pre AI, a shift in the MSP market to what we'll call MSP 3.0.

Speaker B: What is it you foresaw? Because that's interesting. I think most people still aren't on this, but uh, yeah, what is it you foresaw?

Speaker A: So managing technology has become somewhat commonplace, right? And our customers have gotten smarter about being able to do it. The software market has gotten smarter about building platforms that automate much of this. So things that were really, really hard in automated managed services provision have become really easy. And so it used to be that the goal of an MSP was to get a customer to perfection. Your servers are always patched, they're always secure, they run 24 7, they're 100% cost optimized. If you can get a customer there, you've done your job right. I think MSP 3.0, uh, that's the starting line, not the finish line. I think that needs to be the ante to even sit at the table. M I like to think about technology as just disappearing. I don't want my customers thinking about technology at all. I want it to disappear so that the conversations we have with our customers are not how many servers did I patch today? And did your backups run last night? But now that you don't have to think about technology now that it has disappeared because it is always on, it's always cost optimized, it's always secure, it's always compliant. If you never have to think about that again now, what can you do? We talked about those things that humans are uniquely able to do. How can we co imagine the next iteration of your healthcare organization so that you can better help people, you can get to them faster, improve their experience, serve more patients, help this rural population that generally doesn't have access to health care. What are the possibilities that didn't exist in the MSP 2.0 days when all we talked about was patching and backups. That shift happened before AI, but AI is a supercharger as we move into this new era and those MSPs that really think about deeply understanding their customers business and whose conversations with their customers are Much less about technology and much more about how can technology supercharge the customer's business. Those are the MSPs I think that are going to pull ahead of the pack in the same way that MSPs and MSP 2.0 automation MSPs pulled ahead of the pack. Right.

Speaker B: How do you. The hard question I think is how do you start to do that? Let's say I can even get to the table with the ante and keep the infra on and do the things well. Um, because this is new like you said, we're co evolving.

Speaker A: Right? Yeah.

Speaker B: Um, what's your take on what do we do now?

Speaker A: Yeah, you have to be technically perfect. Right. So you can't make mistakes. And the way to do that is to automate every aspect of service delivery. Um, anytime a human finger touches a keyboard, there's a likelihood of a mistake or there's a possibility of a mistake. But if it touches the keyboard the second and the third and the hundredth time at a certain point, statistically you're going to make a mistake.

Speaker B: Yes.

Speaker A: When you build automation to do those things, mistakes don't happen. It just happens the same way every single time. Predictably it's logged. So removing the human from those toil based activities that are error prone, eventually you automate enough of the service delivery that the technology just disappears. And when you have when that becomes not your finish line but your foundation on which you help your customer build. It's a dramatic shift in the conversation. To me the hard part is not perfecting the technology. We have tools and techniques. We've been, you know, as an industry, we've been doing this for decades. The harder shift is retraining your team and the culture inside of the MSP to stop focusing on blinking lights and you know, tickets and that sort of thing and to get to know their customers and to live inside of their customers industry and to be an intrinsic part of their customers solution team. People who are not people who are used to like swinging tickets and you know that that's a big retrain and sometimes that's a restaff and that transact cultural transition is the biggest barrier I think to this new MSP mindset.

Speaker B: Mhm. Yeah. It's interesting because for so many years we've said we have a VCIO offering and vciso and we want a seat at the board. Um, but really what we were aiming for was just to have great infrastructure.

Speaker A: Yeah, right.

Speaker B: Keep the light on, keep um, it secure. Uh, and what I think is amazing doesn't mean we know the answer, but it's an amazing opportunity. Is now for the first time, we can literally participate in outcomes.

Speaker A: Yeah.

Speaker B: And you know it. For most businesses, even if it's good and helps us do a little more, a little faster computers.

Speaker A: Right.

Speaker B: It's still mostly viewed as a cost center.

Speaker A: Yes.

Speaker B: Um, and my IT budget is as little as I could make it often.

Speaker A: Yep.

Speaker B: But if you're able to help me see more patients.

Speaker A: Yes.

Speaker B: Or expanded to another market.

Speaker A: Yes.

Speaker B: Or actually impact revenue or margin or ebitda, then the budget is pretty immaterial.

Speaker A: Yes, it's right. Exactly.

Speaker B: Roi. It's whatever it takes.

Speaker A: That's the right observation. And so, uh, when we deliver our mbrs and our qbrs, that's where we focus. So the material we send ahead of time, how many tickets did we solve? Are you patched? Are you secure? Where are your compliance gaps? But the conversation that we have is how many more patients did you see this month than last month? And that is a game changing aspect of how we've shifted relationships with our customers. Because the C suite doesn't care about how many servers you patched. Right. I mean your IT director does. C suite cares about what they care about.

Speaker B: Mhm.

Speaker A: And each customer, you know, there's commonalities, but each customer is unique. And so this requires that you really understand each unique customer and their particular mission and how to attach technical metrics to their clinical and business metrics. And that's where AI has been a significant enabling factor. Because doing that research by hand and really getting to know your customer inside and out takes a tremendous amount of diligence. And we've automated much of that. So we have tuned AI, um, LLMs and models to automatically continuously surface those metrics as opposed to the traditional MSP metrics. So our teams that are customer facing always have at their fingertips the right information to have a C level conversation, not an IT level conversation. M And so that's where this move toward MSP 3.0, which has been in progress for quite some time, has been dramatically helped with the use of agentic AI.

Speaker B: Mhm. I want to talk about the cultural aspect because it seems like the value in knowing how to fix a switch or even be deeply skilled in solving tickets, it seems like knowing more about GitHub or uh, being better at being in front of the client and getting the real business value right. It seems like the things are shifting, the skills are shifting. Um, how do you lead your team through that, um, so that they can, especially in this Unprecedented pace of change. Uh, adapt and grow skill sets so that where they are now fits, but where things are going in the future.

Speaker A: Yeah, it's the billion dollar question.

Speaker B: Right.

Speaker A: Like I said way back in this conversation, technology is not the hard part in this, in this transition culture is the long pole in the tent. And you know, it starts again having the right team, intensely curious people.

Speaker B: Mhm.

Speaker A: People that are not fixed in their ways, but their curiosity drives them toward continuous innovation. You need a team of people that is okay sitting in ambiguity sometimes and a team of people that relishes just right. There are people that hate change, there are people that tolerate change, and there are people that thrive in change.

Speaker B: Yeah, I think that's a really, really underestimated point because I think it's easy to go. Well, my team doesn't hate change, but if they just tolerate it.

Speaker A: Yeah, yeah.

Speaker B: You know, I mean, I could tolerate a root canal when I have to.

Speaker A: Right. But you don't want to have a root canal every day of your life.

Speaker B: The less of those, the better.

Speaker A: Yeah, exactly.

Speaker B: Yeah. That's very different than relish.

Speaker A: Yeah, yeah. And you really need that pervasive attitude of people m. Who bore very easily.

Speaker B: Mhm.

Speaker A: Um, to be able to tolerate the dramatic shift that's necessary to come into this new world that I think is really. It's like this is going to be what MSPs need to do. And the MSP 4.0 is going to come much faster than 3 came around and MSP 5.0 is going to come faster than 4. And so again, this will evolve from a manual transmission to an automatic transmission to a continuously variable transmission where you're not even shifting gears anymore, to no transmission. Absolutely.

Speaker B: Like an electric.

Speaker A: Yeah, exactly, exactly. Right. But leading teams through, you know, I can't say that I've done it perfectly by any means. I've made every mistake in the books. Um, and any knowledge or experience that I have in this field is only because I learned, like, say I have a PhD from the school of hard knocks.

Speaker B: Me too.

Speaker A: Yeah.

Speaker B: Yeah. What, what do you think this is,

Speaker A: um,

Speaker B: you're going to do kind of for the shape of like, is it MSP 3 or 4.0? Some people call it Managed Intelligence Provider. It seems like the people and what they work on will be different. How we engage will be different.

Speaker A: Yep.

Speaker B: Is that going to change the business model, how we deliver, how we charge? Like what, what do you think is going to shift next? And in that pot, in that shape of how we do business?

Speaker A: I don't know, that's a, that's a great, that's a great question. Um, I mean we, we will have to continually update our pricing models and you know, where, how we charge our customers. Um, the idea of charging hourly or per incident is so antiquated that like I, I don't know how MSPs stay in business that way. Cost plus that, you know, those models are out the window. Um, today it's most common that MSPs, particularly cloud oriented MSPs are charging kind of some percentage of the cloud spend. Um, because as organizations get more involved in the cloud they become used to consumption based pricing. Um, much of what MSPS used to do has been supplanted by SaaS platforms that are generally consumption based. Um, and so consumption base is I think the standard model today as we drive more toward discussing value received from customers. Did you see more patients versus did we patch more servers? Um, does pricing shift to a value model where we quantify the monetary value of value received and get some percentage of that? I don't know. We're going to have to continue to get creative. Um, the most important thing though is that we have to align our pricing models with what's best for our customer.

Speaker B: Mhm.

Speaker A: So MSP fees are generally viewed as a tax in the same way that IT departments are generally viewed as a cost center.

Speaker B: Yeah.

Speaker A: I want to see us get to the point where a customer is glad to write a check because they've received such tremendous value that that check is a tiny fraction of a percent of the overall value that they've received. Where the relationship between MSP and customer stops being adversarial. Like did you meet your SLAs this month? And it starts being a true partnership where the MSP is a part of the customer's team delivering actual value in a way that the customer can see it and recognize it and value that relationship. That shift will be a tremendous shift and that goes right along with at the same time that the internal customers see their IT departments as creators of value, not consumers of dollars.

Speaker B: Yes, Entirely different mindset.

Speaker A: Entirely different.

Speaker B: I am concerned that m some of the MSP industry we will. What's going on with AI may happen, uh, to us instead of with us. And to me. One of the best examples is SaaS. For the most part they just went and bought the CRM marketing directly. No conversation, no consultation, no security, no governance, just getting the tooling. And it's great that it was maybe priced on consumption or priced, you know, it could deliver value quickly. Yeah, but For a lot of us, we just weren't in that conversation. We weren't really aligned to do so.

Speaker A: Yeah.

Speaker B: How do you think we need to align to so this. So. So the impact of AI and as quickly as it's moving that this doesn't become SaaS for us as an industry.

Speaker A: Yeah, it's a great question. Um, we need, we, we have the opportunity today to get in front of that because many of our customers are coming to us saying, hey, you guys seem to be a little more advanced in your AI adoption than we are. Can you help us, you know, become more like you in your maturity and your tech technical savvy around AI? We need to be that guide point or that guidepost for our customers today so that their adoption of AI happens through us, not despite us or around us.

Speaker B: Mhm.

Speaker A: And that means that we need to be better than our customers are at AI today, which means we need to make deep investments, both financially as well as time and the right people to do the AI adoption because they're excited about it and they're curious. Um, but we need to be providing value around efficient and effective use of AI at our customers today so that we naturally become the resource that they bounce ideas off of. If we aren't in that position, they're going to go adopt AI despite us.

Speaker B: Mhm.

Speaker A: And that will result like you got, essentially the sassification of AI that further erodes the value that we provide to our customers. They can go do AI on their own and then AI begins to supplant a lot of the work that we do. We become disintermediated and ultimately lose value with our relationship with our customers. So yeah,

Speaker B: I, uh, think that the hard truth is a lot of us got into this when it was hard to keep the infrastructure on.

Speaker A: Yeah. Right.

Speaker B: And that was our value add.

Speaker A: Yeah.

Speaker B: If you don't shift your mindset, you'll be the only one left trying to get the blinking lights on when the value's no longer there.

Speaker A: Yeah, yeah, that's exactly right. We need to evolve. Our market has evolved, whether we like it or not. I miss writing code, but I can guarantee you that other than very special circumstances, I probably will not write another line of code the rest of my life.

Speaker B: Right.

Speaker A: Like the world. Whether we want to admit it or not, the world has changed. And it's our choice whether we choose to change along with it or to try to cling to the old ways. M and at Cloudbicity, we know that what's right for our customers is that we evolve. And as hard as that internal change management is, and as deep the investments are that we have to make, it's our responsibility as our customers partner to make that internal shift so that we can continue to provide the value that enables them to most effectively serve their patient.

Speaker B: Mhm.

Speaker A: Yeah. Yeah.

Speaker B: We can't forget who they serve, right?

Speaker A: Yeah. And you know, as much as I love technology and it's really cool to talk about agency loops, and we can never forget that the constituent behind the scenes is always that patient that is frightened and sick and needs help. And that is the reason that we do what we do. Unfortunately, what we do day in and day out is really cool and it's fun and it satisfies our curiosity. But we can never forget why we do what we do.

Speaker B: I love that. Jerry, for anybody that would like to connect with you or pick your brain, uh, if you're open to that, what is the best way they could find you?

Speaker A: Yeah, so you can see Cloudticity. It's a strange name. So that's how it's spelled. Cloudticity.com. my personal email address is Jerry G e r r yloudticity.com I answer every email that I get, so feel free to reach out anytime.

Speaker B: What a, uh, generous offer. Um, thank you for being on MSP mindset. This has been an amazing conversation. From culture to agentic loops to business model. Um, and uh, very few people are as deep in this as you are. Um, this is, this has been amazing and I've learned a ton. Thank you Jerry.

Speaker A: It's always a great conversation. Thanks for having me, David.