AI as a Service: Synthreo’s MSP Playbook (EP 935)

Hello friends, Uncle Marv back with another episode of the IT Business Podcast here at IT Nation. Day 2 at the Rosen-Shingle Creek Resort. And I'm going to try to filter out the noise, folks, but they decided to put refreshments right outside my booth.

So everybody is walking by, grabbing drinks and snacks and waving and distracting us, but we're going to push on here. I do need to say thank you to Thread for being the presenting sponsor here. And right now I am joined by Callan Sapien with Synthreo, and we chatted with them last year, I believe, because of PitchIT.

But saw you there at the booth. Don Sizer said we needed to chat, so here we are. She's amazing. 

She's just absolutely amazing. I just kind of do what she says, and I'm really glad to be back and be able to continue this conversation. And, of course, I did ask her, Mike, why do we need to chat? And she's like, oh, he's a really good person and they do a great thing.

And I'm like, all right. You know, I think something that Dawn shares and the tech degenerates in general with something that's core value of ours, which is being that go-giver. You know, I'm a student of Arlen. 

He's my mentor. And I don't know if anybody knows Arlen Sorensen, founder of HTG and then Evolve. And so I hope that it shines through that we, my team, myself, care a lot about this industry, this community.

And I've grown up in it from a young adult through now an older man. You know, so it's very important for us. You're calling yourself an older man now, huh? I am. 

I crossed that 40 threshold. And so now things hurt for no reasons. All right. 

Yep. You're in the club. Real quick, before we get off on a tangent here, let me go back and make sure everybody knows who Synthreo is.

Yeah, absolutely. You guys started in 21? So we technically started in 19. We acquired the company through partnership with a man named Ivan Svac, who had built a machine learning AI automation platform before LLMs existed.

And so when LLMs existed, he incorporated that in. Like many technologists and researchers, he took some money and burnt through all of that money. And we met him in a weird happenstance way and convinced him not to just sell his company, but partner and continue on his vision.

So we acquired the company in the start of this year, finished the close, and then adapted and enhanced some of the things to make sure that it was MSP ready. It was already multi-tenanted, secure. So we officially launched in the summer.

Okay. All right. Because for some reason, I got this whole 21 after the year of our COVID is kind of when you guys burst onto the scene.

Yeah. I was working on automation. I was working on AI.

So that's probably where all that comes together. Okay. Yeah, yeah.

All right. And specifically, is it true that you're a no-code platform? Yes, but as someone who has now built three no-code platforms, there's a lot of code on the backend. And that's actually something that's really important to us is we are as much a services company as we are a SaaS company.

And we lean into that. I don't care what the VCs say. Our mission is to help MSPs deliver AI as a service through hook and crook, right? We have a technology that we feel is very powerful and scalable and no-code, low-code.

But we also have a set of services to build and offer go-to-market because everybody's trying to figure this out at the same time. So our job is to make people successful through both the services and the software. So now let me ask the question that may ruffle some feathers.

It seems as though the last 12 to 18 months, this whole idea of AI agents and MSPs having to sell AI, copilot and all that stuff, it seems as though this is the time where we've just been told, boom, you got to go now. Absolutely. You were doing this before.

So what's the landscape for you feel like right now? It is a story of have and have-nots. There's a few people that get it. And I think there's a few people that have been disrupted by AI integrators or by AI as a service companies, and they've been surprised, right? And so I think they've lurched forward.

But the vast majority of the people that I talk to actually struggle with the how do I find the problems? Like for me, our company's mission statement is there's a marketing one, but at the core of it, it's I want to find problems that MSPs and their customers are having. And so what MSPs are coming to me about is they're saying, I really want to do this. I understand some of the technology.

I don't understand it, but I know that we can get there. But I don't know how I can apply this to my customers. And so they're leaning into things like AI chat.

And when I look at the cycles that we've had over the last decade, I look at cloud, 365, it made us $100 billion industry. It also took our margins from 70, 80% project work, block time, down to whatever Microsoft would give us. 16 if you're lucky.

16 if you're lucky, right? A third of the industry unprofitable. And then I look at security, this promise of high margin. And then what do we do? MDR, SOC, all these things.

And so AI, I see the same thing happening again, where we've got all this high margin work here that we don't necessarily know how to unlock because it's a business conversation, not an IT conversation. And then we're leaning into the IT conversation, which is how do I secure chat? How do I do that? Which is a margin of maybe $10, $15 a seat. And so it's not needle moving things that we're focusing on.

Well, it's not just that, it's how do we, what's the phrase? The horse has left the barn. Absolutely. How do you corral it back in and stuff? It's like we're being asked to put guardrails on stuff that is growing too fast.

I mean, our customers, this is the weirdest thing. And it's not just AI, it's all tech. We're at a time where our customers are getting to see stuff and use stuff before we are.

And for two and a half years, it was done. We told a lot of our customers, stop, pause, we don't know what's going to happen. Stop, pause.

And what did they do? They did it anyways, because it was accessible, to your point. And this is the first time that they've led us kind of by the nose. And so I completely agree with you.

And I think where we need to get back to is having a conversation about risk that we've avoided through the security process. We've secured through brute force, people have still gotten breached. So I think that over the last few years, people's relationship with risk has changed.

And AI has enabled that change. And we almost need to embrace it, right? We need to say, look, you shouldn't put proprietary data, sensitive data in an LLM unless you know the source, you're hosting yourself, all those things. But maybe let them take a little bit of calculated risk while you build out a business that allows you to offer things like AI-assisted workflows, agents, traditional workflows, map everything within their business so that you keep them busy and understanding of how they're going to consume things.

That's one of the ways that I've seen that actually works kind of as a gating mechanism. Give them some of the risk, right? Give them Claude and Gemini. Give them Hats and 3L, right? Give them these things that are going to keep them associated and help them learn how to interact with it while you get the needle-moving, high-margin things ready to go.

All right. But how do you address security from an MSP standpoint? I mean, you sound like you're having conversations, but you're not ready to tag somebody? Yeah. So I would say there's two things here.

I've been a CISO. I believe very deeply in security, right? It must start. But security exists to balance the business risk, right? We're all risk-takers at the end of the day.

And all of our customers are. I was talking to an end customer for an MSP. She owns a coffee shop, and she can't afford a car, right? But she can afford her lifestyle, but she can't afford a car.

So when the MSP goes back to her with, you're going to get breached, you're going to lose your business. I can't afford a car, so I can't do these things. So some customers will always be like that.

My argument would be they are never going to have the money to pay to be able to protect themselves. And so there's an illusion of privacy and an illusion of security that we give them. We have to figure out how to enable without risking our own thing.

We need to focus on the ones that do have the resources to be able to do that and get the things in place. But here's my question. Yeah.

A lot of people say that sort of stuff, but they can't afford it. They just choose not to. That's fair.

Or what you're offering, they don't see as valuable enough to need to purchase. I actually agree completely with that statement because I do think it comes from, for her, riding a bike to work, even in the rain, is not enough pain to motivate her for cutting something out. Maybe letting go of the employee.

I think that does go back to the core conversation, which is we've focused so much on the management of the technology and the implementation of the technology that we haven't focused on the value and the pain that it's supposed to be taking away. The profit margins it's supposed to increase so she can afford a car or afford to pay for a pen test because she is a target. Because the moment that the thing happens, she's, why didn't you tell me? I'm losing my business, all of these things.

But I think AI has put us in a position where we do have to take some of those calculated risks and realize they're going to take them with or without us. Right. And we'll potentially be held accountable for that.

I'm not saying let them use the free CoPilot or any of those things, but maybe if they can't afford the enterprise version, don't hit them too hard. It's those types of things. CoPilot is going to exist.

CoPilot is not great, but it's in Word. It's in Excel. We can't escape it.

Every MSP needs to start with CoPilot if they have any Microsoft in their environment. And I say that as a competitor to CoPilot. Right.

But here's the other side to that. Because these companies are putting these things in their hands without our help, our assistance, our guidance, our blessing to some degree, doesn't that make our job that much harder? It does. It absolutely does.

But what I will say is there is a session that I've been working on, and it's essentially, I had done it before, which is turning shadow IT into revenue. Turning shadow AI into revenue is a tremendous strategy. Understanding and identifying what's being used and then showing people how to use things correctly.

I'll give you an example. My product has a front-end AI chat. We use everybody else's LLM to process it.

We have built-in RAG, all these things. But the way that I use AI chat to be additive is like a team. When I need something, I invite my team to a meeting.

I invite Gemini. I invite Claude. I invite Manus.

I use Threo. I use ChatGPT. And we collaborate together, me moving things through the way that I would have with my team.

When I do it that way, I'm able to show an end customer how not to ask what I should make for dinner, how not to ask to generate this email for me. I can build a three-page press release in an hour and a half that's cited and optimized and all those things because I'm using the different personalities. And that's just chat.

That's not talking about magentic and things. But MSPs diving and experimenting and doing more than what their end customers are doing by signing up for one license of Copilot or ChatGPT can help teach tricks and get into the business that will reestablish themselves as the experts. And then they can talk to things.

Practice with prompt injection. Practice with model poisoning. All of these things that are ways of opening up that conversation.

Master the tricks and then master the attack vectors to be able to go in and really take back the expertise. That's kind of what I end up advocating. Okay.

I thought you were going to go a different way when you talked about turning shadow IT into revenue. Oh, no. Yeah.

I was thinking you were talking about using AI to go out and find what's happening in a company's network. Oh, that is part of it. And then saying, hey, look, this is happening on your network now.

Here's where you're vulnerable. And where your opportunity is. Okay.

Because it's very easy to see on the network. At LionGuard, when I was product officer there, one of the things we had built was a way of scanning. And I worked with Avik on this, John Harden, also a major AI advocate like me.

We had ways of scanning the network looking for AI use. Not just chat GPT and chat, but also within the products that we've deployed. Because every product is incorporating AI into their stack.

And so it's being used whether or not the end customer even knows they're using it. Absolutely. And so you can find ways of, oh, I can secure this.

Did you know? Go to that customer. Did you know that this product, Adobe, is now 40% of the workload is AI? You don't know that you're even interacting with it. And you're uploading your documents into that cloud.

Absolutely. Absolutely. And it's being treated.

Interesting. And then that's a great conversation starter that's more traditional to what we have used in the past. I do think we do need to develop internally or hire the skill set that goes to the customer and says, what's the process that you wake up and go, shit, I don't want to go to work today.

What's the thing that you hate? Do you know what your employees hate? And then sit there and listen for a little bit. We were just on with a partner of ours. Sometimes I put on the shirt and I'm like, hey, I'm MB Systems.

I'm Callan. I'm an AI expert. And we just sit and have a conversation.

There was a car wash with 21 locations that has a cashier. And that cashier also sends emails. That cashier also answers the chat and all these things.

Takes her away from her job. And so we just went in and like, what's your pain? My pain is my cashier can't upsell. I get a line of 10 people because she's answering the phone.

Can you guys build something for us and with us that just answers emails and knows our catalog and tries to upsell people? Like, yeah. And that car wash is willing to pay $1,000, $2,000 a month for this thing so that their employee can just go do their job. The employee's not consuming AI, but now the company is.

And then MSP facilitated this game creation pain reliever. And just by having some of those conversations and understanding the tech stack that's out of what we're managing. They don't know the point of sale system that's being managed, but we figured that out.

And it's those types of conversations that we need to be able to have as well. All right. So we talked nothing about your product.

Is there anything at least before we head out here that you want people to know that if they come to you, you can help with specifically, anecdotally, and start to make money in this thing? And that's the key, right? We have to be sustainable. For me, I'm sitting here. This is my 14th IT Nation.

I was an early employee, sub 100 of ConnectWise. When I was, we were about 5 million in ARR. When I left eight years later, thousands and thousands of employees, almost half a billion in revenue.

That existed because we empowered MSPs to grow. Our company is designed to capture that again in the AI era, automation era. We have a platform that has front-end consumption.

We have a low-code, no-code, full-code builder. And then we have the services to make those work in your client's environment, including go-to-market. You're not in this alone.

If one of us needs to jump on a call, do training, we're here for everybody. I mean, that's really it. We are that abstraction layer for your MSP to connect to all the things AI.

Not compete with ChatGPT, not compete with Anthropic or Google or N8n, but actually take these things and get them in their place. Okay. Have you embraced the MIP, the Managed Intelligence Provider? You know, Pax8's a great partner of ours.

We're going to be in the marketplace. We work with Microsoft. I think that the strategy is sound, right? I'm still, I am a healthy skeptic in all things, even though I embrace risk and run whole hell head in.

I don't know that changing the name necessarily creates a new market. It may be controversial to say, but for me it's services, and we need to get back to focusing on business services and delivering business value. All right.

We'll see if anybody notices that and does a pushback on the MIP word. Yeah, yeah. All right.

Well, Callan, thank you for stopping by. We can now both go and tell Don this was done. Sounds great.

Chase us down. I want to drink Nomalort. I'm good.

What is that drink? She was trying to tell me. I couldn't hear. It sounds nasty.

It's horrific. So in Chicago, turn of the century, not this last century, but the century before, an immigrant from Sweden came over and brought a wormwood aperitif. It's like absinthe if you strained it through flaming golf balls and cooled it off in hot gasoline.

Oh, that sounds lovely. That's a wonderful way. And if you ever go to Chicago for the first time, I'm a Detroiter, grew up in Detroit, went to Chicago in my 20s to visit a friend.

They have this thing called a Chicago handshake. It's a shot of Malort and then a shot of Old Granddad, and it's a horrific start to any night. Neither of those sound like— It's bad.

I mean, why would you do it? Is it a badge of honor? It is a badge of honor. It really is, and there are some people that claim to like it. We drove around with Shawn Lardo at Evolve.

Don't tell me Lardo liked it. No, he did not. But we went around at the golf tournament for Evolve.

Golf cart, alcohol, one of the things I bought was a bottle of Malort because, you know, it's an initiation. Okay. Eighteen people took a shot.

Eleven of them either liked it or said they liked it. I don't believe any of them, but that was eye-opening. I thought you were going to say eleven of them puked because that's what it feels like.

No, I expected it. And there's somebody—Jason Slagle's running around here with Malort jerky. If he offers you jerky, don't take it.

Don't take it. Oh, hell no. No.

And jerky's such a wonderful, pure thing. Listen, I'm not a jerky fan to begin with, so my wife has been watching these off-the-grid shows. We watch Naked and Afraid.

They all talk about when they get a kill, they got to turn it into jerky so it lasts longer. Salt it. And I'm like, okay, in that sense, maybe.

Or if you—what's that? Buc-ee's. If you go to Buc-ee's and they've got the wall of jerky. That makes sense.

Right. But something that just sounds putrid. And it breaks apart and gets stuck in your teeth.

It actually starts out tasting pretty good, so you lose it. Dawn, fortunately, saved me yesterday and gave me a glass of wine, so I just swallowed the whole thing. So somebody said that drinking water afterwards makes it worse.

It does. You start to, like, mouth water because it's a liqueur, not a liquor. It actually, like, activates and thins it out.

It's not great. So you said it gets stuck in your teeth. Is it like a sludge? That was the jerky.

It was the jerky part, right. The jerky turns to sludge. It was.

It was. Not again. All right.

So that just reminds me never to go out after dark with you folks. That's true. The Tech Degenerate group is, you know, look, is degenerate always a bad thing? Well.

We can argue, but tech is. Listen, there is the phrase that, you know, sometimes doing the right thing ain't doing the right thing. Absolutely.

Nothing wrong with that. And some of the best innovations in this world have come out of people being degenerates or going off course. That's okay.

Yeah. But. We were told not to mess with that apple.

And then what did Newton do? Sat under that tree. Yeah. But drinking that nastiness.

Don't do it. Doesn't sound good. Don't do it.

Yeah. Yeah. Malortuitas, it's turned into a whole thing.

If there's a signal chat for IT Nation, and it's about every other meme is Malort. I mean, you know something's good when the tagline is, tonight's you fight your father. All right.

We're going to go ahead and end the show. Alan, thanks for stopping by. Thank you.

Colin Sabian. Thank you so much for having me. Synthreo.

AI powered solutions for your MSP. That's it. Thank you, sir.

Thank you very much. All right. That's going to do it, folks.

We will continue more with IT Nation and have better, pleasant drinking experiences. We'll see you soon. Holla.