Listen to the podcast
Read Transcript
Coming to you live from Pax8 Beyond in Salt Lake City, Erick and Rich discuss the many ways Pax8 is trying to accelerate the path from managed service provider to managed intelligence provider, plus three things you can do to reduce recurring issues that consume resources, frustrate clients, and suppress profitability. Then they have two interviews for you, one with Rex Frank, Pax8’s VP of Academy, about making the journey from MSP to MIP and another with Elliott Hyman, CEO of Lyra Technology Group, about what being an MIP looks like in today’s market. And finally, one last thing: Proof that it’s possible the person flying you home from your next business trip isn’t a licensed pilot.
Discussed in this episode:
The Pax8 Managed Intelligence Provider Playbook
Some guests on this podcast are clients of Channel Mastered. Compensation plays no part in their appearance or the content of the discussion unless the episode they appear on is a “bonus episode” explicitly labeled as sponsored.
Transcript:
Rich: [00:00:00] MSP owners, let me ask you a question. What would you do with an extra 10 hours every month? Reality is that too many MSPs are still wasting valuable time on manual invoicing, payment reminders, and chasing overdue accounts. And according to a US bank study, poor cashflow management contributes to 82% of small business failures.
Discover how leading MSPs are automating their entire accounts receivable process, reducing days sales outstanding, accelerating collections, and making it easier than ever for clients to pay during the next live MSP Mastered webinar on June 18th at 10:00 AM Pacific Time. Stop letting billing slow your growth.
Register now at bit.ly/arautomation. That’s [00:01:00] B-I-T.L-Y/arautomation,
all one word. And now, onto the show.
And
three, two, one. Blast off, ladies and gentlemen. Welcome to another episode of the MSP Chat Podcast, your weekly visit with two talking heads talking with you about the services, strategies, and success tips you need to make it big in managed services.
My name is Rich Freeman. I’m chief analyst at Channel Mastered, the organization responsible for the show. I am joined physically side by your other co-host, our CEO and chief strategist at Channel Mastered. His name is Erick Simpson. Erick, tell folks where we are.
Erick: We are in beautiful Salt Lake City, Utah, Rich, at Pax8’s Beyond 2026 conference
Rich: And we’re playing a little sleight of hand timing-wise with you.
The conference was actually yesterday and the day before as we record this. In the middle of the show, you’re gonna see some interviews we did during the show with people who are attending here, and you will get a feel for the hustle and bustle. This was actually a [00:02:00] pretty enormous conference. It’s very quiet in here right now.
Yesterday at this time, there were about 3,500 people milling around. Big show, a lot going on and maybe we should just dive right in.
Erick: Shall we?
Rich: Let’s. So a lot of announcements coming out of PAX8 here at the show this year. If I was to tie them together, they basically fit under two themes, I would say.
And so first of all, a year ago at this time, PAX8 introduced the concept of the managed intelligence provider. This is their vision for how MSPs should, must evolve going forward in the age of agentic AI to bring the power of agentic AI to their customers. It was a vision then, and they talked a little bit about an agent marketplace that was gonna help with that.
We’re 12 months down the road, and a lot of what they were talking about here at the show now was bringing that vision to life in two ways. There’s some infrastructure we might not have time to get into, but it is very interesting. And then there are some enablement resources, and [00:03:00] that’s where I really kinda wanna focus in because if there was a messaging theme around what they were talking about at the show here, it is what I would call the need for speed.
And so the context there basically is gonna be familiar to MSP chat listeners and viewers. We spoke a few weeks ago about how Anthropic is spending billions of dollars now on a services business specifically focused on mid-sized companies. It’s not because Anthropic aspires to be a services giant.
It doesn’t. They’re under incredible financial pressures right now to drive consumption of Claude, and we, the channel, is not moving fast enough for them, so they’re gonna do that services work themselves. You’re gonna hear later on in the show about how these AI implementers, AI consultants, non-MSPs, non-solution providers are talking to SMBs right now.
They’re probably talking to your customers whether you know it or don’t, and you’re gonna hear later on in some of the interview segments that SMBs aren’t waiting for the MSP to call them and suggest some AI work. The [00:04:00] customers are reaching out to the MSPs asking for help with AI. So this market is moving very quickly, and you as an MSP can’t afford to dawdle as these things are going on, and they are well aware of that at PAX8.
And so a lot of what they’re doing is aimed at accelerating this journey from MSP to MIP, which as you will hear later in the show, is not a quick or easy one necessarily. So Two big announcements. They launched what they’re calling the Managed Intelligence Provider Program, and this is specifically a training and enablement program for people who wanna make the journey from MSP to MIP.
It, and it’ll guide you through that in a customized way based on your vision for what you, where you wanna be playing in the AI economy going forward. And then as a bridge measure, because again, it’s gonna take time to make that transition once you decide to make it, as a bridge measure, they introduced something they’re calling Managed Intelligence Services.
This is delivered through the Pax8 professional services arm, and basically, [00:05:00] it’s a white label set of services that you can bring into your customer environments today. So while you are training up the workforce and getting yourself to a place where you can do this kind of work on your own, rather than wait and watch AI implementers and Anthropic, take over your accounts, you can get in there now and start delivering value for your customers, and retain and deepen the trust.
This was a big thing they talked about at Pax8, the trust that you have with your clients right now, that they want AI, they are being offered AI by all sorts of people who aren’t you. They would prefer to get it from you because they know you, they trust you. To retain that trust, you need to get moving.
The Managed Intelligence Services piece of what they were talking about here is gonna allow you to get moving immediately while you are training and doing that hard work of preparing to really rebuild your business around this MIP concept.
Erick: Yeah really interesting announcements this week, Rich.
I’ve never [00:06:00] seen a vendor or a marketplace or a distie deliver a service that actually will do the work for the MSP for their customers. They will help, with pre-sales, they’ll help close the opportunity, they’ll help deploy, they’ll help manage, maintain, and this ride-along mentality that gives the MSPs that extra boost of confidence that they need so they’re not figuring it all out themselves.
Following the model, the MIP model that Pax8 has created and is delivering through Pax8 Academy and then actually helping them win deals early on. And, Rich, that’s how MSPs learn new things, right? That’s how we learned it in our MSP. We brought in some subject matter experts when a client wanted something new that we’d never done before.
We did a couple of deployments with them and then figured out how to, fly ourselves and then take it to market and expand it. So it’s the timing is now, and I think you said, Anthropic isn’t waiting [00:07:00] for the channel. I think Pax8 is identifying that and getting partners prepared to hurry up and meet that opportunity and demand.
Rich: Scott Chasin, the CEO of the company gave a really interesting, I thought, keynote on the opening morning of the show, and he walked onto the stage to the sound of a ticking clock, and that, that was, th- setting up the tone for his presentation and a lot of what we got here, and that, that is the story.
There is a massive opportunity for you out there, and the clock is ticking on it. I want to say just a few words about the infrastructure piece I talked about, and then we’ll move on. So they a year ago, Pax8 says, “We’re gonna build this agent marketplace.” That is now known as an agent store. It is live as of this show.
And so th- this is a place where you’re gonna be able to go and quickly acquire and roll out third-party agents created by vendors. They also announced something super interesting called the Agent Gateway. I wrote a post on Channelholic, my blog, about three weeks ago, where I was talking about how MSPs [00:08:00] are really gonna need to get good at token management and token optimization and AI spend optimization.
I wrapped that up by saying, “Man, I wish there was a multi-tenant tool for that.” That’s what this Agent Gateway thing is. So y- you’ll start out a- acquiring agents through the agent store. You’ll use the gateway to scale and manage those agents. And then coming later this year is something that they’re calling their agent c- orchestration and commercialization platform, which is where if you, in your evolution as an MIP, decide to build agents, you’re gonna be able to roll those out to multiple clients simultaneously, manage those, and monetize those, collect the the money that people are paying for consumption of those agents through th- So th- there’s a sort of step-by-step progression that they’ve set up for folks who really go all in on on building agents into their business.
Erick: Yeah. What’s really unique is that multi-tenancy capability, where you can create your [00:09:00] agent, say, for a specific vertical, specific roles within verticals, and then begin deploying those to other like customers in those verticals and in those different roles, and then manage them from a single pane of glass, which is what MSPs like to hear, Rich.
Rich: Absolutely. Now the reason you need to think about token management and optimization is because not every token is priced right, and it turns out not every ticket costs the same amount, which brings us to Erick’s Tip of the Week
Erick: I wonder if we should make an announcement about all this noise we’re hearing in the background.
Folks, we’re, yeah, we’re in the expo hall where- while it’s being torn down. That’s how dedicated we are to bring you this week’s MSP Chat podcast episode. So hopefully we can fix some of this in post, but if not, bear with us as they tear down everything that you can’t see that’s happening behind our camera right now.
So let’s see if we can struggle through, [00:10:00] Rich. So the tip of the week I wanna talk about today, Rich, is a fundamental service delivery tip of the week. When we’re delivering service desk or NOC services or even, SOC services, we’re gonna get alerts that create tickets. Tickets are addre- addressed by the ser- by the delivery team, whether that’s the service desk, NOC, SOC what have you.
And when we have what we call a recurring incident in, in the language of ITIL framework service management an incident is a business interruption. The job of the tier one or level one service desk team is to address that and resolve that service or business interruption as quickly as possible.
User issue, something down, something like that. When an incident recurs, that now is escalated into what’s called a problem in ITIL, in an incident management methodology. So I wanna talk about the problem tickets, [00:11:00] okay? Because this is where MSPs can, without tracking it properly, can really start eroding customer satisfaction, profitability, even employee morale, things like that, when we just can’t seem to have a process where we’re addressing these problems.
So if we have a recurring incident, this is now, maybe it’s the same user, maybe it’s jumping around between different users, and we have to find the root cause of that issue. That absolutely should be taken out of a level one a tier of a service desk and escalated higher, where someone can spend the appropriate time to identify the root cause of that recurring incident and solve it once and for all.
So three quick tips on how to get your hands around this and how to make sure you’re not dragging down your level one technicians chasing something that seems to be just, tr- troublesome and hard to fix. Get that off of their plate. The job of a service one technician is to close as many tickets as [00:12:00] possible, and those are incidents that are these minimal kind of business interruptions and things like that.
So identify, go through your, ticket history, and if you categorize your tickets properly, issue type, subtype, and item, so that you know exactly what the general area of the issue is, what the device or the user that was impacted and what the actual issue was, and then identify those recurring problems and try to figure out what those are.
And get a workflow going immediately so that tickets that look like that in the future don’t go to level one. I always talk about tickets that should always go to level one and tickets that should never go to level one. If some- something that seems like we’ve seen it before, let’s not have somebody work on it and then escalate it out, and then go to the bottom of the second queue right?
And then it takes longer to get it resolved, ’cause that’s also a problem. If you find something like that, assign an owner in your [00:13:00] organization that is the permanent fix owner. Who sol- who figured that out? Who solved it? Was it Erick? Okay, Erick, if we see anything that smells like this ever again, we’re gonna send it to you because you’re the owner of that.
See if that fix is still good, and if not, adjust the approach, and then document that so that we can continue d- delivering that in the future. And then measure recurrence rates. How often do these problems occur, right? We’re tracking and trending not only what types of tickets and incidents we’re receiving, but what problem issues we’re receiving.
And sometimes that causes us to look at maybe the solutions that we’re deploying, maybe the hardware we’re using, maybe the services we’re managing, and make different make different decisions about how we deploy and manage these moving forward. We may rip and replace something that’s too problematic
Rich: It’s interesting the a twist on that advice came to mind as we were talking about that there.
Because to an increasing extent already, and this is only gonna increase from here, the L1s are bots.
Yeah. [00:14:00] And and so there are a lot of questions out there in terms of what … there will still be humans in the help desk process. But what kind of work will the- they be doing?
And so the stuff you’re talking about, the root cause analysis and remediation, that, that’s maybe the stuff the L3s and the L2s down the road are really gonna be focused on. And then the other thing that oc- occurs to me is that the AI, it might be a while before the service desk automation tools are really good enough to do the actual root cause analysis and remediation themselves, but they’re very good at identifying patterns.
Erick: Right.
Rich: So in terms of, you know- Yeah … figuring out what are the recurring issues that are problems, and then even who is the person who should be the owner, because- Ah … they’re always most efficient. Yeah.
Erick: That’s a great insight, right? So basically you’re feeding all of that ticket information into AI, having AI do the analysis, find the trends, find the things that stick out, and then leverage it to identify the [00:15:00] proper resolutions, to identify the owners that identify those resolutions, and then help you along the way.
So you’re getting 10 times more work out of an AI than it would take a level three engineer juggling, fireballs all day long to try to, work on that.
Rich: All right. Folks Erick and I are going to take a break. They are dismantling a very big show here, as you can hear. And because it was a big show, we’ve got not one, but two interviews for you this time.
The first one, which y- you’re gonna get right when we come back in here from the break, is with Rex Frank. He’s the Vice President of Academy, the head of Pax8 Academy at Pax8. There’s really no one at the company who i- is more immersed in the actual work an MSP has to do to get from MSP to MIP.
He’s gonna talk about that with us in a few moments. And then after that interview, we’re gonna be joined by Elliot Hyman. He’s the CEO of Lyra Technology Group a 120 MSP strong organization thinking in some advanced ways about AI. So we’re gonna see what it looks like to be an [00:16:00] MIP today from him.
That’s all coming your way after the break.
And welcome back to part two of this episode of the MSP Chat Podcast, coming to you live from Pax8 Beyond in Salt Lake City, Utah. This is actually the first of two spotlight interview segments we’ve got for you here, and we’ve got the man of the hour for you here as we talk managed intelligence provider at this event.
His name is Rex Frank. He is the VP of Academy at Pax8. Rex, thanks for joining us.
Rex: Oh, I’m excited to be on with the power couple here. This is amazing.
Erick: We’re the power trio now,
Rex: Rex. Power trio.
Erick: Say it right.
Rex: Boom.
Erick: Yeah.
Rich: Rex, obviously the, this power duo, we’ve known you for a long time, but for anyone in the audience who is not familiar with you, just tell them a little bit about who you are and what you do at Pax8.
Rex: Oh my gosh I started in the industry in ’88. I think that’s 38 years or something like that. Grew up on the technical side of the house into the [00:17:00] service manager role. 2010 I founded a company called C-Level Operations, patterned off of what this guy did at MSP University. He probably would never admit that.
Erick: Now you’re getting to classic rock days now.
Rex: Yeah. But just coaching MSPs on getting their operations and finance in order. And then Pax8 acquired C-Level five years ago in May 1st of ’21, and we started building Pax8 Academy, and it’s been an amazing journey helping partners. Nick Headey really convinced me to do the acquisition because we were measuring the number of positive, impactful moments we were making with partners in the hundreds, right?
And N- and Nick said, “Man, I’m gonna help you measure that in the thousands.” And that’s… We made that happen. It’s been awesome.
Rich: So obviously a huge theme of what Pax8 is talking about here, has been talking about with partners for a year now, is making the journey from MSP to MIP, managed intelligence provider.
In your [00:18:00] experience, from your point of view, is that a journey that any MSP can make with a little help maybe from a Pax8 or somebody else if they commit to it? How hard is that? Can anyone actually pull that off?
Rex: How I… First let me start with 2024 was a rough year for me personally, right?
I’m not gonna ask you guys how old you are, but I’ll share. I’m 55, right? And I have been in this industry through, what is it? Five or six major industry transformations, and I just wasn’t sure if I was in to do it again. And I the more I’ve talked about that, the more partners talk to- come up to me afterward and share that they’ve been on a similar j- journey, right?
So 2024 I really wasn’t sure if I was gonna retire or go all in. But I did know I either had to hang up my hat or go all in that there really shouldn’t be an in between. And I think a lot of MSP owners are struggling with that same decision.
[00:19:00] January of 2025, I was at Pax8’s LKO, our leadership kickoff in January and Scott Chasin just inspired me.
And in that moment, I decided I’m all in, and it’s been a kind of a crazy year and a half since then.
Erick: So Rex You’ve been in the industry for a few decades, as have we. We’ve seen this journey that MSPs have been on, right? From the inception of transitioning from break-fix professional services.
We did that stuff. And then trying to, figure out, subscription-based managed services, flat fee, pay us in advance. Getting past all those objections, getting clients to… the customers to say, finally say yes.
Rex: You’re giving me some PTSD here- … but keep going.
Erick: Don’t worry, man.
You still got a few more good years left in you. And then, the cloud thing, right? And the challenge of trying to get clients to understand, And I think MSPs at first too is, we wanna, we want to touch and see the servers. [00:20:00] We’re selling licenses, we’re selling boxes, and all this thing into the cloud.
This whole evolution every decade or two it’s something different. And then security, cybersecurity, which is, the overarching, priority for all of us delivering services to customers, and then trying to get customers to, say yes to that. And then all of a sudden we get hit by AI.
So through all of your experience, and talking to MSPs and working with MSPs- through this entire journey. What do you feel are the bottlenecks the challenges that MSPs are now facing with having this conversation with themselves first as kind of customer zero on this AI journey? And what guidance do you give the partners you’re working with that are trying to figure out, how quickly to adopt and when to adopt, trying to figure out the conversations to have with their customers, bringing stuff [00:21:00] to market.
We’re still in a very, the Jell-O’s not frozen yet in the refrigerator- Yeah … a thing, right? So I know I just hit you with kind of this, here’s your life. Now we’re faced with AI. How are you coaching and guiding MSPs through this journey now, and where are the friction points?
Rex: Yeah. Let me put a little c- context. It- y- you brought two main kind of points to mind for me.
One is, I started in ’88, but the… IBM introduced the PC in 1982. And really all of the 1980s was about inventing how business would use a PC.
And then all of the ’90s was about integration projects and networking and hooking all this stuff together.
That’s a 10-year span. And then we realized we created a little bit of a mess that needed to be managed. And as we went into the 2000s, we were working on the managed services, like how do we manage this thing.
So and that was pretty much all of the 2000s and well into the 2010s.
And, but there’s a 30-year invention, [00:22:00] integration projects and managing that.
Rich: Right.
Elliott: But we’ve done that in… with AI. We’re going into really the second year, right? Like-
Erick: Yeah …
Elliott: like we, we were still in- It’s compressing.
Erick: It’s
Elliott: compressing … we’re still in the innovation stage.
Erick: Yeah.
Elliott: And we’re gonna do the integration stage next year.
And in, in ’28 we’ll be well into the managed stage, right? So three decades compresses to three years. Crazy. And then all of those business curves, like the S curves that you were talking about we’ve successfully managed the curve jumping from each of those. And the AI or digital labor curve, as I probably wanna prefer to talk about it is right in front of us.
Erick: Right.
Elliott: And that was actually kinda confusing at PAX8 to, we… In January of ’26, Scott Chasin said, “Our mission, or our vision is to power every small business with the digital workforce.” And we were all like, “What?”
Erick: Ooh.
Rex: But he set our vision on the next curve, right? And that, that [00:23:00] was kinda confusing to us.
But once I realized, oh, he’s causing us to start that leap onto the next curve- … in, in January of 2026. But the other thing that, that is on my mind is the Gartner Hype Cycle, right? We start with the innovation trigger. Yeah. You rise quickly to this peak of inflated expectation.
Erick: When does the trough come?
Elliott: And then you dive- … right into the trough of disillusionment.
Erick: Yeah.
Rex: And then if you’ve got a product that is, that the industry really wants, it’s gonna curve up into that slope of enlightenment, right? And these AI technologies right now are, like, diving into the trough of disillusionment as we’re trying to figure out how- How they’re actually gonna work in business.
And if you haven’t seen the Gartner AI Hype Cycle, check it out, ’cause they’ve plotted it every year since ’22 all the way into ’26. And and there’s some predictions in there that, that show agentic AI really not even hitting the top of the [00:24:00] peak of inflated expectations in 2025. It’s already halfway down into the trough of disillusionment.
By 2027, it’s moving it into the slope of enlightenment. And and so all of this is coming fast, right? And then all of those curves that you brought up that we went through from break-fix all the way to cloud to security, as an industry, we led all of those.
But we- we’re really not leading AI.
It’s coming at us sideways, right? The, the- … the products that we put into our client sites, the software applications, they’re getting the AI features. It’s being driven from, to us sideways, so we’re trying to adapt. We’re trying to make the decision, how do we figure this out? In the meantime, it’s happening fast and we’re in the innovation stage, but we gotta be ready by next year to- Sure
to start doing integration projects and doing it in a way that we can manage it. And I feel the burden in academy where, PAX8 is building the a- the agent store and we’re [00:25:00] working on MIP program to help our partners make this transition. But because the monetization is really gonna just start now, but be- becoming big i- 12 to 18 months is when we’re really gonna see true monetization happening.
But it’s on my shoulders in academy and my team to deliver the preparation so that we’re all ready for it when the hype cycle says we’ll be in the slope of enlightenment.
Rich: And Scott Jason walked out on stage yesterday to the sound of a ticking clock, right?
Rex: That was impactful.
Rich: Yeah.
Rex: And then they brought it back right at the end, too.
Rich: Yeah. Yep. Yeah.
Rex: Tick.
Rich: So let’s get, j- to the degree that, that time is limited to make the journey, it’s a journey that you have to commit to, as you did recently. Th- that’s because there’s a lot going on. It’s not like you just learn what this AI stuff is, and you’re good.
There, there are technical dimensions, business dimensions. Let’s start with the technical ones, though. Let’s almost get those out of the way a little bit and just talk from the standpoint of a typical [00:26:00] MSP, what are the new technical skills they’re gonna need, and how difficult is it gonna be for them to acquire those?
Rex: The tech is the part I’m least worried about. I often talk about how being in our industry space, especially on the technical side, when you learn something today, it’s gonna have a shelf life of three, possibly five years, right? So on a five-year shelf life of anything technical we learn you have to relearn like 20% of everything you know every year to stay viable in the industry.
And so we’re already expert at learning new things. That’s, that part does not freak me out.
Rich: Yeah.
Rex: The part that freaks me out is realizing that it- it’s almost like an HR question, because we’re used to leading humans And I’m having to learn how to lead humans, but also lead a digital workforce, right?
The agentic world that we’re coming into and developing the… You heard [00:27:00] that we’re… Scott Chasin coined this term last year, right? “Lead with humanity, scale with intelligence.” And he- it’s really tying the humanity to the intelligence which is digital labor. And the we did a workshop within the Pax8 leadership team to start to identify what AI-centric skills are we gonna be looking for in our hiring process?
And I was like, “Wow, our partners have to figure out what are the AI forward skills to add to our recruit- to their recruiting process as well.” That, that part’s a little bit scary. But then to have the business process and business outcome discussion, I think that’s probably 10X on top of what our VCIOs are capable of now, right?
So that’s a skill set that we need to continue to develop. I think an interesting thing that I’ve experienced personally is as my career’s progressed, I was super tech curious for the first 20 year- 25 years of my career, but [00:28:00] as I got more business curious and helping business owners run their businesses, I convinced myself that being tech curious was a bad thing, and I needed to focus personally on becoming business curious.
And the last year, w- you know, once I made this decision, I’m in on this AI thing-
…
Rex: I’ve had to almost reteach myself how to be tech curious- … to use the tools and that’s been a journey for me. And I think a lot of our partner MSP owners are gonna have to go through that same transition to become tech curious again.
They’ve been relying on their teams to do that.
Erick: It’s funny, Rex, I was just talking to an MSP a few minutes ago, and we were comparing notes on, how we’re each adopting AI and things like that. And he made a comment to me that immediately I, doubled down on.
And he said you know what? It’s interesting because it’s re-energized me because now I’m learning this stuff.” And I said you know what? I’m gonna, I’m gonna, agree with you. I’ve been so engaged with vibe coding, and I got this from Rich. Rich has been doing it.” And he’s And [00:29:00] I’m like, “Really?
You’re getting that kind of results?” And then I started doing it, and now I’m switching LLMs and doing all this stuff. I’m up at 2:00 in the morning because I’m not tracking time. And just in it, right? Doing it. So you made the great point about it re- re-energizing our curiosity from a technology perspective, but you also mentioned that you’re not that worried about the tech part of this journey for MSPs today.
And you mentioned about building that business and consultative acumen, because this kind of a conversation with AI is all about business outcomes, and it’s a lot different than what I think traditional MSPs that are still in their m- maturity journey with their technology are starting to really become aware of delivering business outcomes to move that conversation into this boardroom from the server room.
But AI it’s a requirement. You can’t just go in and sell, licenses and SKUs now. You’ve gotta [00:30:00] build a consultancy, a practice, and you’ve gotta build the training for the team to be able to have these conversations at that level. So what are your thoughts on that?
Rex: I’ve had the opportunity to talk to some of our most AI-forward partners, MSPs and their leadership, and it’s a common denominator I’ve been discovering, where the owner who’s divorced themselves from doing the technical work, has decided that they need to lean in, go back to their roots when they started their business doing the technical work.
And when they went back to their roots to develop the vCIO, right? The owners were the vCIOs until they could figure out what it meant and so that they could bring somebody into the role. They’re having to step back into that technician role again.
To lead, to figure out what it is so that they can establish the business model so they can bring a team in to, to manage it going forward.
I’ve seen that common denominator happening.
Erick: So that’s a customer zero kind of a perspective. Yeah. But then also, it’s the vC-A-I-O role [00:31:00] development, right?
Rex: Yeah. More letters. Awesome.
Erick: Yeah.
Rich: Progress. We’re making progress.
Rex: Yeah. No more TLAs, three-letter acronyms.
Rich: So one of the the business skills that an MSP has to cultivate on the path to becoming an MIP and becoming a vC-A-I-O is the ability to assess a client’s needs.
If y- you don’t just come in and say, “Here’s our AI service. Sign on the dotted line.” And you guys have worked out… just to give people a feel for the kind of conversation they’re gonna need to have with end users. You guys have worked out a six-step process basically,
Rex: Yeah, so the MIP playbook that we released in October talked about the six plays, right?
Discover and then sell, and then there’s a fork whether you’re gonna be a broker reselling others, somebody else’s agents, or a builder building your own agents. And then it comes back together into the implementation phase and then the management phase, right? So then that’s the six plays in the MIP playbook.
And so as we [00:32:00] started out to build the MIP program we so far and this is gonna continue to build, but we’ve started heavily with the discover play, and that, yeah… And we mentioned managed intelligent services that, that Pax8 has been building. That’s been out of our ProServe organization, which I also lead in Americas.
And we put a job rec out for a business process consultant and, it was my first time really looking for a Six Sigma Black Belt that was AI forward and tech curious to help us build build what does that discover play really look like.
Erick: Was it hard to find that individual?
Rex: It… there was a lot of applications but it, we really pretty quickly narrowed it down to just a few that had they wanted to be a teacher at heart. That was a big component of it.
Erick: I was more curious about the AI forward- The AI forward … component of the requirement.
Elliott: Yeah, I am very fortunate.
We have a amazing talent acquisition department at Pax8 [00:33:00] that I… I don’t see them until they’ve narrowed it down to a few for me to talk to. But they were able to get that done.
Erick: Awesome. So you talked about the, the broker seller and the builder innovator persona. For MSPs in the audience, map those to, like existing roles that they know they have now, and then how do they get…
How do they close the gap between those roles now? So a broker seller in today’s MSP that’s not thinking AI would be who in that organization? What’s that persona? What’s that role?
Rex: Unless they’ve had a DevOps role, they probably don’t have a builder component, right? Because if you’re gonna be a builder some new skills for MSPs is gonna be around understanding a software development life cycle.
And, put the product management side. Yeah. I think MSPs have made a lot of progress in understanding that the services that they deliver are a product and wrapping go-to-market around that. And, as a service product that they’re delivering. But they… [00:34:00] I think there’s, th- there’s a lot of learning to happen around go-to-market with a development offering.
The SAS lifecycle folks are very familiar with this, but I don’t think the managed services community is currently very familiar with that. Doing patches and the management aspect and managing the security changes that are happening. Like th- this is a muscle that if you’re gonna be a builder, that the MSP’s gonna have to hire in or build or acquire or something.
Erick: So there is an inherent- expertise in any of the folks like in a NOC or a service delivery department. It has to be someone with kind of more of a development, a background- Correct … to understand that life cycle because it is more of a SaaS-
Rex: Yep …
Erick: a perspective. That’s interesting.
Elliott: Part of your decision to go all in is deciding that you have a way to fund and invest in building that part of the business if you’re going to be a builder. And that’s a big part of why Pax8 developed the agent store, right? Is because we know m- majority of the MSPs are gonna go the buyer [00:35:00] route.
That’s what I got. Or the, I’m sorry, the broker route where you work with agents that already exist. And, that’s why Pax8 is investing so heavily in the search capabilities of the agent store because we know the builders are gonna w- wanna expand their marketing activities through the brokers in the Pax8 ecosystem.
And and then the matchmaker piece is really coming into its own right now to match the brokers with the builders and procure that through the agent store. A- and the, the discover though and the sell is probably happening a- as the MIP is making that transition.
And the s- the discover, the new skill set is discovering business processes and then going through a rating system to decide which processes are most aligned with a digital workforce that have to build an agent for it, and then work on those first where you can get the small wins early. And then the sell, right?
It’s a much higher [00:36:00] level, deeper business process, business outcome discussion, which is a skill set that I think, yeah, most MSPs are gonna have to build within their organizations. And then, the integration thing, I, I think we’re gonna see the engineers that are AI forward probably pretty easily stepping into that role.
The NOC teams as we get into the manage the tools are coming. We saw Microsoft just released- The, … Oh, what do they call
Erick: it? Agent 365.
Rex: The Agent 365, right? The
Rich: multi-tenant.
Rex: With multi-tenant- With the … Ooh … management.
Rich: Yeah.
Rex: These tools are coming at us- … and solving some of the problems that have been an impediment to getting into the business, so- they’re coming at us fast.
Rich: You mentioned managed intelligence services before. T- talk a little bit ab- for folks who are brand- it’s a brand-new idea, so everyone in the audience is probably brand new to what that is. Talk a little bit about what that is, but also the role that or something like that can potentially play for an MSP making the transition over to a managed intelligence provider in terms [00:37:00] of partnering with somebody else to fill gaps to provide functions that you don’t have natively
Rex: in-house.
Yeah. Pax8 has always been partner first, right? Like John Street has a history with MX Logic of being partner first. N- never sold direct, and that, that is in the core DNA of Pax8. So we, I think that’s a, that history and trust that we are partner only is one of the things that’s gonna make us successful with this, is, and all of my enablement programs, the premise is if we can help the partner get into a new line of business, right? We’ve been doing it with Azure forever. Maybe they don’t have an Azure skill. Our professional services organization will help the partner do their first couple of Azure projects, and we even do ride-alongs to help their engineers come up to speed.
And eventually we’re happy for, to see the partner start doing their own Azure projects, right? Like I’m not trying to build something to do all the projects across the industry. But it’s more of an enablement role to help them get going, and then we see their Azure [00:38:00] spin, start to hockey stick.
This … We know it’s gonna be the same thing with this agentic stuff. So you know, I’ve got my team of engineers working on not only building agents, but also learning, making sure we’re ready to help deploy the agents that are in the agent store. Because if we can provide that enablement for a partner to get into the business doing their first few or as many as they need then they’ll- Yeah
it’ll help to transition that adoption of selling digital labor and managing digital labor
Erick: Yeah. So Rex, some of the announcements here at Beyond 2026 included even deeper kind of offerings than I was expecting from a managed intelligence service. Understanding that MSPs are looking at this AI opportunity, from a fresh start perspective, and how MSPs typically crawl, walk, run solutions, into [00:39:00] their customer base and expand them over time.
Talk a little bit about how, the crawl stage can be perceived for Pax8 partners, and then the walk stage and the run stage with your managed intelligence services. You hinted at it right now, but peel back the ending a little bit because, in our experience as former MSPs, we know that, boy, it’s hard for us to, develop and take a product to market when we’re doing it on our time, on our client’s dime, and we need all the help we can get.
So I was very impressed with what you guys the, pro- promoted today or this week during Beyond with a new service, it take… And I appreciated the thought that went into it when I was asking specific questions about how are you actually gonna help partners get there?
Rex: Yeah.
Urgency is an interesting thing, right? With AI coming so fast, we are f- we are feeling the urgency to get out in front of things and try to be as visionary as we can [00:40:00] and prognosticate what things are gonna look like a year from now. And, we’ve got Scott Chason, who’s amazing at understanding the technology so deep that he can be pretty accurate i- with his visions.
And and, we have to deliver, we have to deliver that. And ye- yesterday in Lane Brannan’s talk, he t- talked about his personal mantra. One of them is, “Scare yourself,”
Erick: right? Yeah.
Elliott: And we’ve been leaning into that a bit, too scare yourself. It reminds me a lot of John F.
Kennedy getting in front of the nation and saying, “We’re gonna go to the moo- the moon in this decade,” right? Yeah. And you got a bunch of engineers in the background going, “How are we gonna do that?” Then-
Erick: Figure it
Rex: out. And it just triggered this- Yeah … massive amount of innovation.
Erick: Yeah.
Rex: And the concept of we don’t really know how to build a spacesuit, right?
But let’s invent the tests we have to do to find out how a spacesuit has to perform, then we can build a spacesuit. And that’s a lot of what’s been happening within PAX8 lately.
And the sense of urgency around being ready to help the partner with this transition it has [00:41:00] been, i’ll go with invigorating as a word for most of my team. It’s been frustrating. I’m not gonna say it’s not without its frustrations but we just feel like we need… we have to be out in front of it and developing of the programs, right? We try to invent something, and then we go test it in a coaching environment, and we get some feedback, and we collaborate with our product and engineering group.
I see Libby walking by. And then then we test it some more, and then we roll out a concept into maybe one of our instructor-led training programs and get more feedback. And then we have a… We had a AI boot camp in April in Chicago, where we took those ideas out to 150 partners and we ended up with a 96 net promoter score on that event, right?
So it’s feedback. This constant feedback, development, test. Whoa, that didn’t work or it didn’t land. Test, feedback. And that’s what’s getting roll- rolled into all of this with this sense of urgency that’s kinda a little difficult to keep [00:42:00] up with. And it, it’s interesting, my academy team is delivering, I think, 38 sessions between academy and pro serve, delivering sessions at this Beyond.
And all of the… Y- marketing wants us to lock in the slides three months ago, right? Yeah. And we’re like, “Not on the AI content.” We’re like, “We’ll be locking those in two weeks before the event.” Yeah. Because it’s all in this constant state of flux, which is exciting, but also kinda creates some employee satisfaction issues, right?
But we’re working through those and trying to deliver on the go to the moon vision that, that Scott gave us, and it’s been a lot of fun. And to see the stuff it- it’s amazing. The Beyond conference kinda creates this universal deadline across the company. And the collaboration that’s going on between all the different groups and, it…
with weeks to go, you’re finding some misalignments and… but everybody just comes together and straightens that out and to see it culminate in the announcement is quite a ride.
Rich: Well, Rex crazy busy show for everyone at Pax8. [00:43:00] We really appreciate you taking some time out to join us here on the show.
For anyone in the audience who wants to get in touch with you learn more about you, learn more about Academy the MIP program, et cetera, where should folks go?
Rex: If you’re not a partner, pax8.com/academy or the pro services side. There- there’s a ton of information about us. If you are a partner in the marketplace, left-side navigation, do Academy and which will actually take you to academy.pax8.com.
And the, there’s the MIP program I think is right on the Pax8 homepage. Enroll start- start that learning journey. It’s- it’s free to Pax8 partners for the on-demand. If you wanna go deeper with- with interacting with- with individuals within Academy in our peer groups or instructor-led training or coaching, there’s some fees associated with that.
But all the same content’s available on demand. Just you’ll take a survey, it’ll create a curated learning experience for you in our on-demand learning management system. And start taking [00:44:00] advantage of it and give us feedback. The more feedback we can get, the better we can make it.
Rich: Once again Rex Frank VP of Academy at Pax8, thank you so much for joining us on the show.
Yeah, thank you. Folks, we are gonna take a quick break here. When we come back on the other side, we’re gonna look at this from the other side of the story. So we’ve gotten sort of the Pax8 point of view. Let’s talk with an MSP about what it is like to make the journey from MSP to MIP. After the break, we’re gonna be joined by Elliot Hyman.
He is in charge of Lyra Technology Group, a very large MSP with direct experience in terms of delivering managed intelligence services. He’ll be with us in just a moment. Stick around. We will be right back
And welcome back to part two of this episode of the MSP Chat Podcast, our spotlight interview segment coming to you live from Pax8 Beyond in Salt Lake City, where we are very pleased to be joined by Elliot Hyman. He is the CEO of Lyra Technology Group. Elliot, [00:45:00] welcome to the show.
Elliott: Thanks so much for having me.
Yeah, excited to be here.
Rich: So we wanna dive into the whole managed intelligence concept with you and being a managed intelligence provider. But before we do that for folks who are new to you, new to Lyra just give them a little bit of background information about both.
Elliott: Yeah. So Lyra is the MSP platform within Evergreen, so maybe they’re more familiar with Evergreen Services Group.
Our group covers all the MSPs, about 120 now that are, that are- been acquired by Evergreen and now operate as part of Lyra. I’ve been with Evergreen since the beginning. Met the founders of Evergreen before we’d acquired a single MSP. They were telling me about it, and I thought it sounded pretty good.
It- it’s worked out beyond what I expected. I’ve been doing it now for, this is my ninth year. So when I first joined went and worked in the first MSP that w- that we acquired. It was a company called Wolf Consulting in Pittsburgh, PA. And I actually led new business sales there for a year and change.
The founder of that business, Lloyd [00:46:00] Wolf sold to Evergreen with the plan to retire. He was looking to do something else. And the hope was that if I didn’t mess it up, I could succeed him as the CEO. That, that happened after a little more than a year, and I spent four years as the CEO of that business.
And in that time, about mid-2021, launched the Lyra brand as the sort of umbrella organization for the MSPs that we were acquiring, and that’s just continued to grow over the last gosh, five years now.
Rich: How many MSPs in Lyra right now?
Elliott: About 120.
Rich: Okay we mostly wanna spend our time talking about customer-facing AI services, but I’ve gotta take advantage of the access to you here.
Last November at the ConnectWise event, I sat down with Ramzy Sayoud from Evergreen, and it was a huge topic then, it still is now. But MSPs were… are thinking about the internal application of AI to automate and increase productivity, and a lot of different strategies about that.
And I remember Ramzy telling me that he really [00:47:00] felt like Evergreen has an advantage because he’s got all these different MSPs to test things with, see what works, and then roll that out more broadly. And the y- your 120-some-odd MSPs are the test beds. So it’s been, whatever, seven months, eight months.
Catch me up a little bit on what kinds of experiments you’ve been running, what kind of lessons you’ve learned. Where are you at now within Lyroid in terms of using AI to automate the business?
Elliott: Yeah. So there’s sort of two streams of work happening. There’s the organic work that happens in the businesses with the local leadership teams and the decisions they’re making to innovate with AI, and the way that we interact with them is to go study what everyone’s doing where we’re seeing progress, and then sharing that out across the 120 MSPs.
So we have internal systems, and obviously people that share the best of that across the company. In addition to that, so just to give a number on that we [00:48:00] added Oh my gosh, I’m gonna blank on the number. I believe it was 2.2 million of MRR last year without any incremental head count which is the first time we had done that.
So we are not seeing head count reduction, but we are seeing a slowing of head count growth relative to revenue significantly as a result largely of those sort of organic initiatives. In addition to that, this year we made a mid-seven-figure investment in a group of selected businesses within the group.
W- we’re thinking of it as, y- you’ll read these articles sometimes that, McDonald’s has five restaurants where robots are making all the Big Macs, and there’s no people working there, and they’re testing these sort of new things. I think the touch screens that you see in a lot of places now, that was how they did that.
And we’re doing something similar with our investment in AI operations. So for four of our businesses now we’ve allocated a, a mi- a mid-seven-figure amount of capital to hiring partnerships a- and development of agentic [00:49:00] workflows and other ways to bring AI into our operations and make us more efficient and make, just as importantly, make the quality of the customer experience better.
Which I think ki- gets lost in a lot of this, is that, we set out from the beginning, we said, “Hey if any part of the customer experience takes even, a half step backwards, that’s not innovation, that’s not useful,” because it’s too competitive out there to do it the cheap way.
And customers have too many choices, so it needs to be more efficient and better, or just better and not more efficient. But it every, it always needs to be getting better using AI. So that’s been really important for us.
Erick: Elliot, the first time I heard the term managed intelligence provider was at last year’s Beyond, where when Pax8 coined it.
I’ve heard it since then elsewhere. What’s your definition of what managed intelligence provider means? And based on that definition, does Lyra win?
Elliott: Yeah. You [00:50:00] know- I’m seeing the beginning of, we collectively within Lyra see the beginning of sort of a bifurcation in the market. And it’s just the start of it, but I think that there will be service providers who figure out this AI stuff and do it in a reasonably acceptable timeframe for customers, and then there’ll be another subset that don’t.
And I think we are at the… I think the important thing to, to reflect on, I remind myself all the time we’re not just in the top of the first inning on AI. We’re, it’s a 1-0 count. The first at bat isn’t even, it’s just getting started. And the way that we think about it is that, we’re gonna be on this journey and that we wanna be on the side of the man- managed intelligence provider.
I gotta give Pax8 a lot of credit for coining that term. Because the reality is that it is going to be essential for our customers to be leveraging AI in order to be [00:51:00] competitive over the next five years. And if you’re a provider that can’t deliver that, y- you’re gonna eventually get displaced or you’re gonna be working with customers that are lower value customers and- Lower
and likely looking for sort of a commodity solution provider. A- and our strategy has always been to play at the top of the market. If you look at our per seat price the sort of quality of our customers we have many customers that are, would be household names which is always fun ’cause, we kind of work in this sort of, I used to say that working MSP is an unsexy industry, but AI’s kind of changed that. Like, all my business friends really wanna talk to me now about technology and AI.
Erick: You’re no longer the IT guy, now you’re the AI expert.
Elliott: That’s right.
Erick: Yeah.
Elliott: Yeah. Yeah. Isn’t that great? And it’s like all of a sudden it’s, what we work on is on the front page of the newspaper or the journal every, every day.
Yeah. Which is a, it’s been an interesting change in the industry. But we wanna continue to be that, that high-end provider and that’s gonna require being able to deliver all of the AI tools and solutions that our customers are gonna expect. And so we’ve been going [00:52:00] along that journey.
Rich: So Lyra and MSPs like Lyra are on the supply side of managed intelligence. Flipping it around a little bit to the demand side.
Elliott: Yeah.
Rich: W- what kind of demand signals are you seeing from end users? H- how many inquiries do you get from customers about that? H- how interested are they in hearing about it when you raise the topic with them?
Elliott: Man, it’s been a- another big change, right? We would bang down our customers’ doors to get them up to snuff on cybersecurity for the most part. And, the biggest demand from the market came when people were started trying to get cyber insurance, and they said, “Oh, okay, I guess I gotta buy this stuff so I can get my policy.”
It’s been flipped. Customers are coming to us all the time, constantly, every day asking about, “Hey what is this? What are these tools capable of? Can you help me?” There’s also a huge amount of organic usage of these tools amongst [00:53:00] employees without business owners understanding or appreciating how much of this is out there and where their company data is going.
So we’re more often bringing that to their attention of, “Hey, there’s risk in your business right now that you’re not aware of, and let’s talk about how to mitigate that risk and how to use AI in a way that, protects your business and you still get to, to leverage what these tools are capable of.”
At the same time- It’s moving so quickly. If you’re a business owner small business owner in, any of the seven countries that we operate in, you’re busy being a wealth manager or a lawyer or, a manufacturer or providing healthcare services, whatever it might be. I spend a lot of my time trying to stay smart about AI, and it’s really time-consuming ’cause it changes every day, right?
There’s new announcements constantly and, small business owners don’t have time to keep up with that. They are generally curious. They want to know how it can help them. But the specific applications is where they come to us and say, “Hey, help [00:54:00] me understand what I could do with this and help me understand a secure way to do it.”
And, every one of our companies is having the AI conversation with their customers. It was, I’ll say probably shortly after we spoke a little over a year ago, I was getting up in front of our CEOs and our… any time I talk to anyone in our companies or, I spend a lot of time going on site to our businesses, just kinda trying to preach the good word of of AI.
And, we’re past that stage. People have been people have come around and been converted to what this is going to do and is doing. And and now it’s about actually implementing this stuff and actually, making impact with it. And in terms of that, we’re just scratching the surface.
We have companies doing really interesting high-value AI implementations for customers, saving real time saving, full people’s worth of time for customers. But that’s just been emerging over, let’s say, the last eight, 12 weeks where we’ve really seen those projects actually close. So there’s [00:55:00] been sprinkling, but where we’re really seeing the momentum pick up is just the last eight to 12 weeks, which feels like a lifetime in the AI world.
But even six months ago we weren’t seeing the volume that we are now. It’s every week we’re seeing more and more companies both start to offer those projects, close those projects, deliver those projects for customers. So we’re in the middle of, we’re in the middle of the beginning of the first inning right now more than we were a year ago.
Erick: Yeah. It’s interesting. I used to say, the great way to get these customers that historically have been reticent to invest in cybersecurity and all the things that you need them to do, they said A, cyber liability insurance policy,” that’s the pull-through, just like you mentioned.
That’s it. That’s it. And but now AI is the pull-through. AI is the way to get these customers to level up their cybersecurity because of the threat and the opportunity and the leakage and the, shadow AI and all the things that you just talked about. So understanding now that [00:56:00] customers are really starting to see the potential for AI- MSPs have been trying to figure out what to deliver to these customers.
They’re still figuring out how to, get the most out of it for themselves, and now they’ve got customers that are ringing them up, just, “Hey, help us with this.” So what are you seeing, Elliot in your experience, the low-hanging fruit, like the easy opportunities to have conversations, to get customers to say yes to something, but then can be expanded over time?
And then what kind of skilling does the MSP need to be able to have those conversations and deliver on the promise?
Elliott: Yeah. So this has been a daily, hourly top of conversation within our team.
Erick: And- Do you have a silver bullet for me? Come on … No. No, I,
Elliott: unfortunately, I don’t have a silver bullet. And like I said, we are
Th- this is happening in real time. Yeah. And I- Could you go sell a meeting [00:57:00] transcription agent to every customer? Sure. I don’t… We’re not as focused on that because we don’t think that’s where the differentiated value will be.
Erick: And then MRR really, no strategic value over time.
Elliott: And so what we’ve been focused on is two prongs of maturing our businesses. The first is making sure that we have a very high quality VCIO having conversations. VCIO is the term we use in the industry. Whatever you wanna call that person,
Erick: VCAIO?
Elliott: Yeah, I’ve been hearing that lately.
Yeah. That’s a new one that’s getting coined. I’ve
Erick: been saying that for a while. Me and Rick- Okay … Spann, I think we oughta trademark that.
Elliott: Have you… You should have. It’s getting… it’s like MIP. It’s like all of a sudden these new terms are- Yes … becoming very commonplace.
Erick: Yeah.
Elliott: So the first piece is you’re having a high quality substantive conversation with customers about their business, right? So
Erick: you a- Discovery conversation, listening …
Elliott: ongoing learning about what they do, who does what, how much of it do they do- Yep … and educating them on how, on what AI even is.
Because, like you said, [00:58:00] they don’t have the time to go do all the work that we can and should be doing. So our job is to go learn and to go bring that back to our customers as education, not as sales. And so that’s piece one, is making sure your account management function, whatever you wanna call that, your, the TLC you give the, your existing accounts, that’s gotta be really sharp, and you have a, have to have a really high quality person having those conversations.
The second piece is, so now you understand their business. You’re having ongoing conversations about AI. You’re asking good questions. Now we’re layering in an actual, “Okay, what are we gonna go do about it?” process. And so what that looks like for us is essentially a consultative but more structured consultative, almost survey-like conversation where we start to understand the flows of their business to help identify what those opportunities are and then start proposing things that, that we think we can deliver for them.
And frankly for our companies, what we recommend is, [00:59:00] most of our companies are starting by going to clients that they have really high trust relationship, long-term relationships with, people who listen to them, right? Everyone kinda… Every MSP kinda has that subset of customers that really listens to them, and that subset that’s sorta “Ah, just answer my ticket and I don’t wanna talk to you.”
So you go to those relationships that you, that, have high trust And we’re doing this work at highly discounted rates to start. We actually recommend that for our companies. Hey, do we… I, so Paul Dippel’s on our board, and I get to talk to Paul he’s an advisor to us, and I get to talk to him, a few times a quarter.
And he always preaches to me through this whole process of “Hey, you’re not good at, you’re not even serviceable at something till you’ve done it seven times.” And then he has now he’s 20-something where you’re really like mastered it. And so our guidance and what our companies have done on their own as well is, give it away at half the price.
It… And it’s actually more expensive work, right? You’re using some of your best technicians to do this work. But when you’re getting started, just do something, prove to yourself that you can add this value, and then [01:00:00] that becomes your case studies that you bring to your other customers. And we’re doing the same thing internal.
I thought Nick did a, Nick Heidi did a great job this morning talking about customer zero. We also talked about customer zero, where it’s hey, get smart about this stuff and start trying things in your internal operations. Yet, yet, if you can do it in your internal operations, now go sell it to a customer at a reduced price.
Let them know. Lay it out for them. “Hey, we haven’t done this a bunch of times before. This stuff’s happening right now, but here’s what we did internally. It’s adding a tremendous amount of value. Do you wanna try it with us?” And, hopefully they’ll have some grace with you at, when you present it that way, and obviously do it for less money, which always helps.
And we’re just building those early reps. But the stuff we’re se- in general across the company, we have a handful that we call our trailblazers or, our, or our more frontier, I guess is now the word in AI- Yeah … that have been doing more for a while. But as we’re spreading it across 120 businesses, that’s what we’re seeing is people starting to get their first handful of reps with customers, implementing a solution.
Can
Erick: you share with the listeners what are the the things that really clients are saying yes [01:01:00] to when you’re… great insight, hey, don’t use this as a money maker. Use it as an opportunity to pay down your costs, learn, develop, and then hopefully build product and package service.
So what are the one or two top things that customers are saying yes to now?
Elliott: Yeah, the- They’re
lookers …
the ones that are most interesting to me are- so if you spend time in SMB world, you know that there are employees in, most SMBs that are doing fairly rote repeat We’ll call it the human plug in the system, right?
Email comes in I’ll give you an easy example. We have a wealth management customer. They receive all these different documents from their clients all the time and someone needs to look at that and figure out who’s the right advisor and how should this be treated and all these sorts of different things.
It’s super easy for AI to manage that process and get the work where it needs to go, and then you can start building simple workflows off of that and says, [01:02:00] “Okay, when it’s a W-2,” I don’t know, make something, simple. Okay, that every time will go to accounting or it will be processed in this way or this information will be added to XYZ database.
That’s work that people have been doing, all over the world at many companies and certainly in SMB land that just doesn’t need to be done anymore. And so things like that where it’s-
Erick: Easy admin stuff- Oh … repetitive low value
Elliott: That’s where we see w- let’s call it the, you know-
I don’t wanna… I was gonna kinda throw a, Yeah, I think administrative is the right way to describe it. Yeah. And there’s people all across SMB land who are doing a lot of that work that is just clicking through screens and moving data from one system to another manually.
That’s just super easy and but customers are, like, amazed by it.
Erick: Right.
Elliott: The s- thing that we’re starting to see more of now is data visualization. So building, And this is why, it’ll be interesting, you obviously have had the SaaSpocalypse this year [01:03:00] and, I don’t see people switching off a lot of systems of record yet or rebuilding systems of record.
Yeah … I’ve seen a little bit of that. A little bit of it. We are seeing a little bit of
Erick: it. We gotta come back to that. Come on.
Elliott: Yeah. D- we haven’t seen a lot of that, especially in SMB land. I know that, there’s conversations about it maybe more at the enterprise level, but, And if they’re doing it, they’re not doing it super well.
I d- we couldn’t talk more about that. But data visualization, there’s so much data in these small businesses that they have-
…
Elliott: That they’re not looking at or don’t have in a usable format. And even for us internally, we are… So we got everyone on the team at Lyra a Claude subscription and a Replit subscription.
And the quality and impact of the output from vibe coding, which I think it might actually be the right now the most important output of this whole thing, more than even agentic workflows. The things that you can produce quickly bespoke for your business using vibe coding are incredibly high impact, and you [01:04:00] don’t need technical skills.
We have, people on our team who are largely financial analysts who are now building software all over the place to solve real problems that they’re experiencing in their work. And so bringing both those visualizations. And really it’s not AI, it’s just software, but it’s better suited to your use, and it’s really easy to make.
That stuff is blowing people’s minds right now because they haven’t seen that before, how quickly they can make We have a process I’ll give you an example for us internally, and this is, We acquire a company, we can ingest all of their vendor invoices look at what products they’re using compare it to our database of vendor agreements.
So we have agreements with a bunch of different vendors across the ecosystem, and recommend to that company, “Hey, you could move your buying of SentinelOne from whoever you’re buying it from to Pax8, for example, and save 50 cents an endpoint. Hey, you can, … You- you’re using this backup product.
We have an agreement with this backup product. We believe that they’re largely the [01:05:00] same or maybe this one’s better, and it could save you $5,000 a month on your backup bill,” things like that. One of the analysts on our team is a generalist a- analyst, built that for us in a few days, and now it’s used across the company and it’s finding tons of savings.
That’s a simple example, but that the software production is really … When customers see that even people internally see it they’re blown away. That that’s been really powerful ’cause it’s not invisible workflow. It’s like you can see the output, and you can see it so fast.
Rich: We’re figuring all this stuff out in real time now, and so I’m gonna ask you a question that requires you to look down the road a little bit, but I’m … I know you’re thinking about this a little bit, and it’s a topic I ask y- relatively newcomers to this field about all the time, which is pricing.
So particularly as you get past the those early wins and the things that are relatively obvious for the end user in terms of where the impact is and so on, and you start automating entire [01:06:00] workflows end to end and you have the opportunity to talk to the end user about how y- you were paying somebody $60,000 a year to do this, and we’re gonna have agents doing it for you, and you just pay us 40 or 30.
Outcome-based pricing. Th- there are a lot of alternative pricing models out there that are very n- new to the managed services world. Are you experimenting with any of that, or do you have any kind of thoughts about where that might land?
Elliott: Yeah. So I do. And the answer’s we don’t know, right? I think, you know- Six or 12 months ago I was panicking because playing out this, these predictions, our customers, historically our customers have grown naturally, organically.
And we get this sort of tailwind on, on growth because if you’re adding 5% to your workforce, you add 5% more seats, and we charge whatever we charge per seat. We’re picking that up every year, every month, whatever it might be. And [01:07:00] for that to reverse, which is what we’ve been talking about and even seeing in our own operations, right?
We’ve flat-lined staff while growing revenue that breaks our sort of per-seat pricing model. You’re seeing this in SaaS as well, right? They’re, Salesforce is worried maybe you don’t need all these seats because you have, a Salesforce that’s 1/10 of the size, and what is that gonna mean for our revenue?
And I felt better when I said, Microsoft will figure this out.” Microsoft’s not gonna let, 20 or 30% of the M365 seats go away. So they’ll solve it, and we’ll just do what they do. We haven’t seen that yet, and we’re starting to see in the data for our companies that their clients are healthy, growing businesses, declining headcount for
We’ve never seen that before, and we’ve been looking for the last 12 months very closely to see when we would see it. We’re starting to see it. And so I think the first thing the MSP needs to figure out is what does that mean for my revenue, right? If all my customers are getting smaller and I’m not getting, at the very least I’m not getting that tailwind I used to get, and maybe I’m actually getting a headwind and we’re retracting, if my [01:08:00] contracts don’t protect me from that’s the first thing that I think you have to think through.
And I think it seems like we’re moving towards some type of consumption usage model, maybe, reselling tokens or something like that and getting margin on tokens. That maybe is somewhere where we’re going. It’s has to play out and become feasible. On the project work you know- We’ve actually spoken to…
So part of our research over the last six months has been talking to a bunch of AI implementers, and I love when they get on a call and they’re like, “We’re expert AI implementers.” I’m like, “Does that mean nine months ago? What, how much expertise can you…” Just because the tools have changed so much, and even if you were an expert a year ago, you might not be an expert at what’s available now.
But we’ve spoken to a bunch of these companies and we have seen the we’ll save you, a million dollars a year, and we’ll only charge you 500,000. It’s gonna be great.” And you look at what they’re gonna deliver, and it’s like a vibe-coded dashboard that looks like a CRM plus a this, plus a that.
And it’s okay I’m pretty sure I can go do that in a week. 500,000 feels a little heavy. I think there’s gonna be a [01:09:00] period of time where there’s, we have another advisor guy named Peter Bryden, who’s a longtime industry veteran who has a line that, where there’s mystery, there’s margin.
And as long as this stuff looks like magic, you’re gonna see people charging all sorts of different prices. Our point of view has always been our- the value is in keeping the customer and doing the right thing for the customer. Technology will change over time. Our thesis has been, for the last decade, SMBs will spend more on technology over time, and if, if we just hang by the hoop, that’ll be helpful.
That’ll work for us. So we’re not looking to fleece customers. And in fact, like I said, we’re… Right now, we’re just learning to do the work well. And I think we’re figuring it out. But the first principle is, don’t take advantage of the customer’s current where they are in their learning process about AI to take more than is fair.
Now, over time, if this becomes rinse and repeat and we’re, chugging through things in 30 minutes we’ll have to think about what the pri- the long-term pricing [01:10:00] is. Right now, we’re just trying to get good at doing the work more than anything else. And I think the monetization, my belief has always been, like, you add value, you can capture value.
And so the big thing is, let’s figure out how to consistently add value in a systematic way where customers are, like, raving about what we did for them. There will be a way to charge and make money on that. That’s our point of view on it.
Erick: Yeah, we talk a lot on the program about the different ways to…
Everyone is talking about monetizing the AI. We just talked about maybe there’s an outcome-based price and something later that… Because, yeah I think that once everybody starts figuring out by coding a little bit, like we’ve been figuring it out lately it is, like you said 500 grand, that seems a little heavy.
We could probably workshop a little bit with our team, maybe bring in an advisor or something in. Do you do workshops, AI, by coding workshops with your customers?
Rex: So we have companies doing all different things. I actually was just at one of our companies called Systems Engineering in Portland, Maine [01:11:00] last week, and they’re doing something really cool where they essentially do a workshop on, on- All of the different uses of AI potentially and how the company could incorporate.
Frankly, right now in SMB world, most of our customers outside of a small number of cities, they’re not vibe coding. They’re not even ready to start. They’re learning to start using Copilot.
Erick: Yeah.
Elliott: That’s the reality on the ground even though- Yeah … we hear about different pockets, but the…
My observation is they wanna talk about it, but they’re just trying to learn how to use the basics. Yeah. They’re not ready or, yeah, they’re not ready for that next step. It’s obviously coming. But we are doing workshops, education for clients all over the place. I thought the program that they had was really innovative, really smart how they’re doing it.
So I wanted to give them a shout-out here. But that’s the s- we’re at the education phase right now. Including educating ourselves. I, I-
Erick: Yeah …
Elliott: I I think the news… There was a great, actually during the keynote today they- Yeah … they made a good point of sort of the [01:12:00] exponential and the inability of the, sort of the human mind to understand the exponential.
Erick: Yeah.
Elliott: And, I think the biggest, and what we’re seeing, the biggest, What is setting the pace on AI adoption is the human ability to adapt and learn, not the technology right now. And so that’s the thing to overcome and work through. A- and, that’s where we’re focusing our attention.
Erick: Yeah. You mentioned that, a lot of these SMBs are ready for Copilot, for a chat kind of interface initially. When, we hear Pax8 talk about knowledge intelligence, we hear a lot about, agents, agentic AI which is that next level from the chatbot.
How does agentic AI play into your strategy? And do you see that as… do a lot of your customers, are they thinking about agentic AI? Are they still in chatbot mode, or are they talking to you about agentic AI?
Elliott: I’m not sure that they’re that different. [01:13:00] The, you go on Claude right now, right? Or ChatGPT and you say “How many people can fit in this convention center?” Okay, that’s one way to use it. You could also say, “Hey build me Create the blueprint for this convention center. Show me three different versions of how we could expand it.
What would the i- impact be on the local environment? How could I make the case to the local government that we should take this plan?” You can do so much there that, with the thinking modes and the extended reasoning and then, all this stuff’s kind of coming together.
And I think there was just a piece of news yesterday that ChatGPT’s gonna totally change up the, their app. OpenAI is changing their app to be bring Codex into- … into ChatGPT, and maybe you bring News in there, and all these sort of things. So I’m not sure that the line that we’ve drawn in the sand between chatbot, agentic, and, whate- It’s all kind of one thing, I think, and it’s just about, “Hey, what’s the problem, and what’s the way to solve it?”
Vibe coding is agentic work, right? You have multiple agents working on your vibe coding.
Erick: Let me [01:14:00] narrow the question then. So creating agents and then orchestrating agents. So w- I was speaking with a an MSP earlier today, and they were saying, “Now, we’ve got to train ourselves.
Agents are like human beings, and we have to think about managing them in the same way, and orchestrating and things like that.” And I hadn’t thought about it that way. So we’re, when I’m speaking about this agentic, it’s like now everybody has their, their chief of staff agent. I have my marketing agent, I have my, my email agent, and my, all this, all these things.
So can we… how are you thinking about that, and how… And are your clients asking for that kind of support yet, or is it something that you’re starting to feel them out with some of these surveys and assessments you’re doing?
Elliott: Yeah, I think it’s largely the latter. For the reasons that we talked about earlier.
Erick: Yeah
Elliott: So I, I think, [01:15:00] I just I think it’s unsettled whether this sort of unit of measure as the agent is ultimately the unit of measure.
I think the consumption is ultimately the unit of measure, so it’s gonna be the tokens or the c- compute- Yeah … seemingly. And you know- I’m just not sure the complexity of, that we think of as agents right now is gonna be the permanent way that we operate with AI.
Erick: Yeah.
Elliott: And
Erick: so- Yeah, I guess in my mind I’m just thinking of it as the next step up, and then the next step up, and the next step up. It’s like that evolution.
Elliott: Yeah another way, like something that I think-
It’s, can we have tasks that are constantly running and working on each other and, can react to inputs from outside the company, like things like that. I’m not sure that I, can pull in tools and do all those things to solve a problem. I’m just not sure that the agent idea is gonna end up resonating as what that is.
It’s just what [01:16:00] computers do now.
Erick: Yeah.
Elliott: And I think we go back to isn’t agent- it was agent first, right? So it was like Salesforce that kind of coined the term and Marc it was Marc Benioff who- … put this together. That was years ago already, and it just feels like so much has changed, and I’m just not sure that unit of measure is gonna end up being the way we think about it.
It’s just like stuff AI can do. That’s what customers, so I’m also not the technical guy. Yeah.
Erick: So just worry about the outcome. I think that’s right. That’s what I want. Yeah.
Elliott: And I think the way we get to outcomes is becoming much more simple and much more approachable for the average person right now, and I think we’re gonna continue.
Apple did an announcement today of, they’re bringing Gemini into Siri. I’ve always said like when Siri becomes as good as ChatGPT Voice- Yeah … you’re gonna see a population of people understand how cool this stuff is- Yeah … that never have before, and I think we’re on the brink of that right now, where it’s just gonna be a lot more blended into everyday life for a lot more people with the way Apple’s integrating it and I don’t know.
I think it’s just gonna feel a lot-
Erick: Let’s [01:17:00]
Elliott: say … more ubiquitous.
Erick: Yeah. Apple can bring it to the consumers, and the consumers then bring it to the workforce a-
Elliott: Yeah, and I I think the voice thing in particular makes it so- Sure … accessible. And Siri’s in everyone’s pocket, right? Or old Siri’s in everyone’s pocket.
But it’s funny because just talking about the ubiquity of AI I don’t know how many times you’ve seen public speaking over the last six months. I’ve been to weddings, go to events, and you see the ChatGPT ticks-
Erick: Oh, yeah …
Elliott: and but it’s a real person that you know saying all the words. But you’re like, oh, I can’t think of one right now, but you-
Erick: But today’s-
Elliott: There’s these giveaways.
Yeah. Yeah. But that’s not the point. Whatever, like the sort of like abrupt-
Erick: Give you three things-
Elliott: Yeah …
Erick: and
Elliott: then,
Erick: Exactly. The tells. The tells.
Elliott: And so we know people are using it, ’cause every time they get up in front of people they seem to be using it. But I’m not sure that it’s as woven in…
I, no, wait. I am sure it’s not as woven into everyday life as it’s about to be with, as we’re catching up over the next six to 12
Rich: months, yeah. Oh, yeah, super interesting [01:18:00] conversation. The key thing that we started out with is how fast everything is evolving. So we would love to have you back on- Okay
Six, eight months from now, whenever the opportunity- Yeah … and ’cause everything will have changed.
Elliott: That’s right.
Rich: And we’ll see where you’re at at that point. For folks in the audience who are as interested in this as we are wanna get in touch with you and wanna learn more about Lyra, where should they go?
Elliott: Yeah. So you can go to lyratechgroup.com. The website is not vibe coded. It was built by me five years ago. It’s not great. If you wanna talk to me, you can email me at ehyman, E-H-Y-M-A-N @lyratechgroup.com. Happy to answer emails. You reach out to me on LinkedIn. Yeah, either one works. Love to talk to anyone interested.
Rich: It just occurred to me. Am I remembering right or are you a New Yorker?
Elliott: I am f- I grew up in New York. I live in Chicago now, but I grew up in New York. Okay. Yeah.
Rich: Knicks fan?
Elliott: Oh, huge Knicks fan. This has been a dream run. I’m very, I woke up this morning like a kid on, Christmas morning because we’ve got the game tonight, game three.
And I’m [01:19:00] just excited to see what Madison Square Garden looks like tonight. Oh. ‘Cause I think it’s gonna be- It’s going to be crazy … wild- Yep … and you know- I was joking with someone just real quick, and he s- he sends me this thing from Twitter or something. It’s game three is the same night as my wedding, and I’m just so bummed I can’t go.
Does anyone wanna marry her? She’s lovely.” And-
Erick: Wasn’t expecting that.
Elliott: Yeah. Nice. Yeah, it was pretty funny. And and, I, and I joked back. I said I’ve been married more times than I’ve seen the… than I, than the Knicks have won the finals in my life.” I’m only one marriage.
That’s right. We all … it is pretty special. And yeah, I’m really excited to watch. Yes.
Rich: All right. Best of luck-
Elliott: Thank you
Rich: very much … to the Knicks- Thank you … and to you. We will certainly be watching as well. Folks, we are gonna take a quick break here. When we come back on the other side, Erick and I are gonna share some final thoughts about this very interesting conversation with Elliot Hyman.
We’re gonna have a little fun. We’re gonna wrap things up from here at PAX 8 Beyond 2026. Stick around. We’re gonna be right back.[01:20:00]
And welcome back. Part three of this episode of the MSP Chat Podcast, where they haven’t come in with a forklift yet and taken us away. But they continue to dismantle the show around us. Two great interviews Erick. I’ve got… there’s a lot we could talk about with respect to the conversation with Rex.
Two thoughts about the conversation that the folks in the audience just h- heard with Elliot Hyman. And one… And this was… This ki- moment came and went in the conversation there, but it was interesting to hear that Lyra has done the analysis and seen that head count is dropping at their customers for the first time.
They’ve never seen that, and these are healthy businesses. The issue isn’t that the companies are shrinking or that they’re not doing well financially. Head count is beginning to drop. And so there’s all this talk in the industry, including in the managed services industry about, per-seat billing, and is there a future for that?
The… Here’s some evidence that, A- and again, need for speed, clock is ticking. This is happening [01:21:00] now, where if you’re billing by the user, by the seat for software, fewer users, fewer seats coming along. So you’re gonna need to find a different kind of model for monetizing. And then the other thing, of course, we…
It was just good to… this circles us back to where we began at the top of the episode. The customers are hungry for AI and relatively sophisticated A- AI based on what Lyra’s hearing from its clients. Those AI consultants are out there, and they are looking to talk to customers who are not getting that kind of assistance from…
So again, it’s not just a future issue. It’s not a theoretical problem. It is a here-and-now issue.
Erick: Yeah. And I think also what was interesting was when Elliot was talking about how they’re pricing these early pi- pilot AI projects. He says, “Hey we’re basically just cover our costs.
Let us get in there. Let us learn,” right? That takes a very strategic, consultative [01:22:00] conversation and a very trusting client because what happens after the p- pilot is successful, then The billing begins at the normal rate. So it’s interesting how you have to have that conversation and almost partner with those customers.
Or maybe it’s a scenario where, this handful of clients is always gonna get this, super awesome grandfather early adopter rate, and that’s just what you’re going to invest in developing your solution portfolio around AI. So you got to think about that as well. I know that, there’s a lot of, conversation always in the channel about pricing and bundling and, it gets back into what you just mentioned about, if the user count begins dropping because of what you’ve deployed, then you’ve got to figure out a different way to monetize it.
And we’ve heard four or five different ways to monetize it, right? By project, by agent, by, by outcome, right? By consumption. There’s all these different ways. So it’ll be interesting, as we continue reporting on, what [01:23:00] AI is doing in the channel for MSPs to see which ones are the winners as we go through that.
Rich: One quick final observation about all of the different pricing models you were talking about, and all of those are being experimented with and debated, outcome, consumption, et cetera. There is a lot less predictability to any of those schemes than there is with the traditional per user, per device ki- you go all the way back to when you were inventing managed services, Erick, and part of what was appealing about it, both for you as the provider and for the end user, was the predictability of it.
I know when I’m … What I need to budget every month- Yeah. … because I know how many users I have and what this guy is gonna charge me for that. And unfortunately that, that world’s going away, and we’re going back to a world in which there’s gonna be less predictability about what to charge and what to pay, and that’s gonna be something for the whole industry to figure out.
Erick: Yeah, you’re absolutely right. That was one of the selling points is oh, you can budget, you can forecast, we’re … you know it’s gonna be the same every month unless something [01:24:00] crazy happens. And now we’re getting back into, this, unpredictability, like you said. And I think it’s just gonna cause the MSPs to just level up in terms of the mature, consultative approach that they take with these clients.
They n- they know that, hey, you got to put gas in your car, right? At least nowadays still, right? ‘Cause we’re not fully, electric or anything like that. But … And gas can go up or down. And I think, as we see these providers of AI begin really optimizing the way that they’re reducing their token burn by moving from, MCP servers to CLI.
We’ve had, some conversations around that where that also can reduce costs overall. But yeah, it’s a moving target. And you know what happens, which, this was another conversation we had recently. What happens when the cost of that consumption does drop, right? [01:25:00] Does the MSP then reduce their rates when they go out and sell these things?
Or do they lower their, the the next agreement will be lower than last year’s agreement? There’s all this stuff that’s happening that we’ve gotta figure out as a channel together
Rich: All right. While there is still time before they kick us out of the conference center here, that leaves us with time for just one last thing, and immediately after we dismantle our operation here and take off Erick and I are both gonna go to the airport, we’re gonna fly home and we are gonna take it for granted that whoever it is who is flying that plane bringing us home is licensed to do that job.
And yet, as you may have heard, because it’s apparently been all over the news recently, there was a pilot at Air Canada for 27 years I don’t even know how many flights he flew in that time, who was not in fact licensed to do that job. He did know how to fly apparently because he was able to, take off and land safely obviously.
Wasn’t actually supposed to be qualified to [01:26:00] hold that job which is a sobering thought if you fly a lot like I do right now. A- and one other weird little thought that kind of popped into my head as I was looking at this, ’cause I… Like literally minutes before we started recording here, I was listening to a podcast where some people were debating the efficacy of AI and the skeptic was pointing out the hallu- it, it hallucinates and so on.
Humans make mistakes too.
Erick: Yeah.
Rich: And this is a perfect… I can… If you don’t know if you can trust the HR function to a HR, just keep in mind the people who are doing that job right now might actually put someone who isn’t licensed to fly into the cockpit of a very large airplane.
Erick: So that would be the variant of catch me if you can-ada.
Rich: Indeed, indeed. Folks, thank you so much for joining us here on the show this week. That is all the time we’ve got for you. We’re gonna be back in another week’s time with another episode of the show for you. Until then I’ll simply point out as I always do, this is both a video and an audio podcast, which means if you are listening to us right now, [01:27:00] but you’d like to check us out on video, go to YouTube, look up MSP Chat.
If you’re watching us on YouTube, but you’re into audio podcasts, go to Google, Apple, Spotify, wherever it is you like to get your audio podcasts. You’re gonna find us there too, and wherever you find us, please subscribe, rate, review. It’s gonna help other people find and enjoy the show just like you. This show is produced by the great Riley Simpson, who’s part of the team with us here at Channel Mastered, where we have an end-to-end family of services that vendors can tap into to build thriving MSP channels.
Channel Mastered has a sister company called MSP Mastered. That’s Erick and his team working with MSPs to help them grow and optimize their business. You can learn more about that at www.mspMastereded.com. So once again, we thank you for joining us. We’ll see you in a week. Until then, please remember, as we always urge you to, you can’t spell channel without [01:28:00] MSP.
Check out more episodes
