Your managed services are hitting every SLA metric and the customer still thinks you’re failing – here’s why

Robert Dutt: Hello and welcome to the ChannelBuzz.ca podcast, bringing news and information to the Canadian IT channel for the last 16 years. I’m Robert Dutt, editor of ChannelBuzz.ca and still your host for the show. Today we’re talking about a problem a lot of MSPs and channel partners are starting to feel, even if they don’t always have a name for it yet, and that’s visibility. As AI workloads, hybrid architectures and distributed endpoints become the norm, network traffic is changing faster than the tools that many partners rely on to understand what’s actually happening inside their customers’ environments. My guest today is Jeff Collins, CEO of WanAware. Jeff spends a lot of time with service providers and enterprise teams dealing with this shift, where accountability for performance, security and uptime is increasing, even as environments become harder to see and harder to diagnose when something goes wrong. WanAware operates in the network and infrastructure visibility space, but this conversation isn’t about the tools, the dashboards. It’s about how blind spots form in modern networks, why they’re easy to miss until there’s an outage, a security issue, or an SLA failure, and what partners need to understand as AI-driven infrastructure quietly reshapes traffic patterns and dependencies. In this discussion, we’re going to explore where traditional monitoring starts to fall apart, how partners can rethink what good visibility really means today, and why the ability to see what’s happening across distributed environments is quickly becoming both a risk issue and a business opportunity for MSPs. If you’re responsible for customer outcomes, but you don’t always feel confident you can see everything that matters, this conversation is for you.

[MUSIC]

Robert Dutt: Jeff, thanks for taking the time. I appreciate it.

Jeff Collins: Thanks, Rob. Thanks for having me on.

Robert Dutt: You’ve been advising partners, MSPs, VARs, these types of folks through a lot of change over time. Why does this moment with the rise of AI workloads and the continuing trend of hybrid networks feel like a real inflection point rather than sort of just the next evolution of the way things look?

Jeff Collins: I think one of the biggest reasons why is because it’s so transformational to what MSPs and resellers and VARs and distributors have dealt with for, let’s say, the last 25 years. If we think about the last major inflection point that they dealt with was really kind of the realm of the hypervisor, this ecosystem where no longer did we have to have a server running an operating system, and that created kind of the whole ecosystem we deal with today. It created cloud, it created containers, all those things were built off this concept of a hypervisor. That was really the last major transformational thing that has happened. Now we fast forward to today and we’ve got this era of AI. We’ve got this era where we’re now taking agentic approaches, generative approaches, to things that our customers deal with every day. When I talk about our customers, those are the customers of the MSP, those are the customers of the reseller, the distributor. Not only are they dealing with that, they’re dealing with this massive evolution in the customer base, but they’re also having to do that same evolution in their own environments. If you’re an MSP and you’re focused on infrastructure, or you’re an MSP and you look more like an MSSP where you’re focused on security, now you’re starting to have to deal with, “Okay, I’ve got these tools, I’ve got these people, I’ve got these agents, I’ve got all these entities inside of my business that are doing something for my customer.” But now I have to think about how am I going to do that faster? How am I going to do that better? How am I going to do that more effectively? Because our customers are getting much more advanced. That’s really one of the biggest things that I see that we’re seeing a lot of, that “Where do I start?” from the channel partner community. When we think about the channel, we know all this stuff is going on, but it seems like such a Herculean lift that I think sometimes it’s hard to know where we make that first step.

Robert Dutt: That makes sense. A lot of this, a lot of AI especially, and to a degree sort of the hybridization of the network, that complexity has come on without kind of a formal network redesign. Like you mentioned the transition to hypervisors and that necessitated rethinking how things were done because it was a physical change. Whereas a lot of, especially with AI, it’s kind of being bolted in, added on as you go. Why does that make the environment today harder to understand than maybe it was for past transitions when you’re sitting there watching it as an MSP or other partner?

Jeff Collins: Well, I think one of the biggest reasons why this era is so much more difficult than the last transition is because we’re not bound by the four walls of our proverbial house. If we think about when we dealt with the last transition, every customer, their physical server sat inside of something they control. So we’ll refer to it as their house because that’s the easiest kind of comparison we can do. In today’s world, there’s certainly a lot that exists in our customers’ houses and in the houses that the MSP or the reseller or the channel partner or whomever it is are engaged in. But so much of that’s going outside of those walls. And when we think about AI, AI is certainly outside of those walls. I mean, we might be dealing with Anthropic, we might be dealing with ChatGPT or Gemini or the thousand other agentic or generative approaches that are out there. Those are all over the place. And now we’re asking these entities to take oftentimes a process-driven approach that they’ve had for 20, 25 years. And how do you change that process-driven approach when you don’t really know where those workloads, where those assets, where that data is going to reside either today or tomorrow, or even if that data that we’re looking at is even going to exist tomorrow. That’s this whole realm. I mean, we’ve been talking about ephemeral workloads for, you know, let’s call it 14 years, 15 years since really the rise of AWS. But now we’re starting to deal with these ephemeral workloads, not just in the realm of infrastructure, but also in data, in generative concepts, in agents. You know, historically, we had Bob Smith, who might have worked in the NOC. Well, tomorrow, Bob Smith is an agent. What does that look like? It’s AI. What did Bob Smith do yesterday? Did Bob Smith, the new agentic version of Bob Smith, did that person do the right thing, the wrong thing, the incorrect thing? How do we manage that? How do we deal with that? How do we process that? Those are all the things that are across the board, just happening at massive rapid scale. And so, you know, it’s a really difficult time right now to be an MSP or a channel partner, but it’s also an amazing time to be an MSP or channel partner. You know, our world, our capabilities are advancing so fast. You think about one of the simplest use cases that’s out there that all of us think is simple, that MSPs deal with every day, is a circuit outage. You know, a telecom circuit goes down and it’s connected to SD-WAN or it’s connected to a router or it’s connected to some type of device that’s out at the prem. And historically, every MSP on the planet’s dealt with it kind of in a similar way. We get an alert from a monitoring system that feeds a ticketing system. It pops up on a tier one agent’s dashboard. The tier one agent looks at it, they verify power, they verify if the router’s operational, and then they open a ticket with a carrier. And then they, and that’s the hurry up and wait type of world. Well, now in the era of AI, that changes that quite a bit, because every one of those things are very process driven. We don’t need people for that anymore. So now we can have a system take that process flow on, do that. Now, historically, we could use a system to do that. We could write automation and a lot of MSPs did that historically, but the problem with automation is automation is static. When we leverage AI, we can leverage enrichment that helps influence that agentic approach. And so now if there’s a nuance going on, let’s say an example is there’s a global power outage. So let’s say there’s a power outage in the entire Vancouver area. We know that. Well, historically, if we’re looking at that, we see all these customers that are down, we might through a tier one agent approach, a person-based approach that following a process, or even an automated approach, not really correlate that. Because if the MSP is in, let’s say, Montreal, they might not realize there’s a large scale power outage in Vancouver, which is thousands of kilometers away. And so when we think about that, that’s really where these things can change a lot from an agentic perspective. And then the MSP gets the joy of being able to repurpose that person to be much more valuable to their organization, that tier one person can become tier two, and that can really start changing that dynamic a lot.

Robert Dutt: Most MSPs would have historically said we have good visibility across what our customers are doing. And probably I would say most believe they have good visibility today. Where does that confidence most often turn out to be misplaced or to start to break down as the model shifts?

Jeff Collins: Yeah, so I would 100% agree that most MSPs, when workloads are static, have great visibility. The problem is that in today’s world, so many workloads are becoming dynamic. And we see that change happening consistently. You know, customers, you know, historically MSPs had problems monitoring services inside of a cloud provider. You have ephemeral workloads, you have workloads that aren’t necessarily a server, they’re much more like a service. So you have things that might be a Kubernetes instance, they might be a Kubernetes runtime instance, they might be a function. Those are all things that are crucial to the operation of a customer. They’ve taken those workloads that historically operated on a machine. And they’ve taken those workloads and now they’re in some type of small form factor instance that exists for a very short period of time. That’s been very difficult for MSPs to deal with across the board. But now we take that same concept and that same concept goes outside of the cloud providers. We now have that moving into inference nodes. We now have that moving into IoT and IIoT and OT, where we’re starting to deal with these ecosystems where these workloads are very ephemeral by nature. They might exist for a short period or components of those might exist for a short period, or the way that those are correlated and analyzed might exist. But if you think about inside of a customer from a business risk perspective, those actually carry the highest business risk. An individual Windows 2012 server has some level of business risk. If it’s running SAP, probably a higher level of business risk. But if it’s one Active Directory node and the customer has 100 machines in Active Directory, it doesn’t really matter in the scheme of the world. And so those are the realities of what happens as we kind of think through this stuff. And so for MSPs, this really drives that visibility gap. You know, we did a survey earlier this year, or actually late last year, sorry, in 2025. We did a survey across the board asking business leaders really what the visibility gap was and what they believed. And we asked business leaders and we also asked IT. It was really interesting to see kind of the dichotomy. When you ask the business what the visibility gap was, it was extremely high. When you ask technology what the visibility gap is, it was really low. Now they were both technically right. And here’s why. So IT was thinking about the visibility gap of the machines that they understand, the machines in their purview. So those might be, you know, an Active Directory server, a database server, maybe you have a web front end. Those are all there. And those are 100% being monitored to that IT team or to that MSP. The problem is, is the business itself is operating on a whole bunch of additional workloads that IT doesn’t necessarily have purview to. And so because of that, we start ending up with this difference of visibility. And that’s why oftentimes when you’ll go and you’ll talk to a customer or you’ll go and you’ll talk to the business itself. And the business is saying, why do we have this MSP who works for us? This MSP isn’t doing anything. And the MSP is coming back with these great reports that are showing MTTR is consistently dropping. You know, initial response time, triage time is consistently dropping. We’re blowing out every single metric that we provided you in an SLA or an SLO. And the business is coming back and saying, but you’re failing. And the MSP is saying, I don’t understand. We are not. And here’s all the metrics. And it’s because of this difference in resources that exist, that is what is happening. And so I think that’s one of the big areas that we always have to think through is, you know, as we’re looking at things and as MSPs look at things, they have to continue to be pushing upward inside of the business to understand all those areas that the business is driving that IT, who they’ve historically sold to, may not know about those resources, especially in a lot of these other spaces, AI, IoT, IIoT, OT, ephemeral workloads, cloud workloads, those types of things that are often outside of that scope.

Robert Dutt: Yeah. I guess when you’re looking at sort of your visibility stopping basically at the edge of the organization, you’ve got all of this out there, pretty significant impacts on real world issues like latency, like security exposure, like the ability to meet those SLAs that you signed up for, those kinds of things.

Jeff Collins: Yeah. Yeah. 100% agreed. And, you know, when you think about the core components that an MSP does, you know, MSPs generally deal with availability and they deal with performance. When you add in the MSSP, now we add in the security component. And some MSPs and MSSPs are more hybrid-based approaches. They may deal with all three. But as you kind of look at those, those core tenant areas have become much more difficult, especially in the last 10 years, certainly in the last year. I mean, the last year has been so disruptive for all that we do. And it’s because those pieces have become much less simple. You know, if I go back 25 years or even 20 years, customers by and large used MPLS networks, rather simple to monitor. You have guaranteed jitter, you have guaranteed latency, you have, you know, all these things that are very easily assumed by an MSP. So if latency exceeds 74 milliseconds between these two individual locations, that breaks the SLA that the provider provides and it’s an easy conversation. You need to go fix this. This is not okay. Well, in today’s world, most of our customers don’t have MPLS networks. Most of them have, you know, sometimes now it’s satellite. They might have Starlink for LEO. They might have 4G or 5G, depending on what portion of the world they’re in. They might have some type of broadband service, fiber broadband, or copper broadband, or some other type of realm. Well, those don’t necessarily have SLAs for that in any way, shape, or form. We may luck out and they have an availability SLA. Maybe it’s three nines or two nines, or maybe not even two nines, depending on what type of service that is. And then when we start moving inside of the network, outside of the service provider, outside of the circuit provider itself, we start moving into other arenas that look like this. You know, historically we had a Dell server, an HP server that had a mean time before failure. Well, that’s pretty easy to understand. If I have a server and it’s going to run for 25,000 hours, it’s easy to understand that life. But when now we’re starting to get services that have an expected failure, and that expected failure is generally measured in less than a year, because the assumption is that the software, the application, resolves that issue. If you’re an MSP and you’re not monitoring the application and you don’t understand the application, you’re now chasing outages that don’t matter. And that’s one of the other things that’s really hard. And we see this all the time. You know, I’ll talk to MSPs and they’re like, “Jeff,” and it goes back to that same conversation we had before of not knowing the business. “Jeff, we get, today we have 30% of our tickets that become false positives. What do we do about that? We’ve gone out and we’ve bought the newest monitoring platform. We’ve implemented AI. We’ve implemented all this automation. We spent $20 million doing that.” These are all real things that I have in conversations with MSPs. And at the end of the day, they still have 30% false positives that they’re working. And the reality is, is because it’s certainly an outage. There was 100% an outage that happened. But the reality is that outage was never going to get restored because the outage was designed. You know, that workload disappeared. A DevOps team or a DevSecOps team deployed a new environment and that workload is now gone. And there’s a brand new workload that you’re not monitoring right now. You know nothing about it. And those are the things that we all collectively have to continually evolve to. It’s that driving up the stack. You know, one of the things that I often see is, you know, we have this proverbial thing that we’ve all dealt with, the OSI model. You know, there’s seven layers to that OSI model. So often in MSPs, we focus on four of them. The problem is, and most MSPs only focus on the first three. They don’t even focus on the fourth one. The issue is, is there’s three more. And those three more are what get driven by the business. And so the more that we can focus on visibility within those three, understanding that, bringing that into our tools, that drives additional value. It also drives significantly larger margin. You know, if we think about margin contribution at monitoring a telecom circuit, that’s a pretty low margin at this point in time. There’s a lot of automation around that. Monitoring a server – that world used to be high-margin, but it’s compressing. Customers are increasingly doing more of this themselves. They’re doing automation directly into their CI/CD pipeline. So it becomes this knife fight. And there’s more and more MSPs that are out there that are also fighting for that same share of market. And so the key is, the more that MSPs can go up market, they can understand, you know, I hate to use this term digital transformation because it literally gets overused every day by every marketing team on the planet. But the reality is, is that if we go behind this marketing abomination of this term, and we actually look at what happens, there’s a ton of value that we can go after. And if we go after that value, and we go after what people are trying to do, we align with that, we can now take those same products, those same processes that we’ve historically had as MSPs, and we can really start evolving that. Moving upward, driving in significant value, taking our tool sets that we may have today, maybe those can evolve with us, maybe we have to make new changes in our tool sets. But the reality is we’re driving that margin upward. So we’re going from maybe our contribution margin to our business today is 30%, let’s say, we can start moving back up into 60, 70, 80% contribution margin from a managed services perspective, which is where we all want to be. We don’t want to be fighting knife fights for 30%. It’s just hard, it’s difficult. Our customer acquisition costs are still generally high. We have salespeople, we have marketing efforts, we have all those things that we’re burning through every day. And we need more and more market share, we need more and more assets that we’re monitoring. And as a result of that, we need better ways that can contribute higher margin and create stickier customers that we’re not in those knife fights with.

Robert Dutt: The situation seems to be putting MSPs in a situation where they’re increasingly accountable for outcomes that they can’t fully see the contributing factors of. Before you move on, I just wanted to double click on that just a little bit and just ask, how does that change kind of the risk profile for an MSP when you’re accountable for those things that you don’t completely understand or have complete control over?

Jeff Collins: Yeah, I would say a lot of that. And one of the things that MSPs have to think through is a lot of that starts at the sales cycle. If you don’t ask the right questions at the sales cycle stage, oftentimes you get pushed into that ecosystem. When you’re looking at the core functional plumbing behind what a customer is trying to do, and that’s the only thing you’re looking at, you often get siloed into that ecosystem. You’re looking at a server, you’re not looking at SAP. One server going down in SAP doesn’t necessarily mean SAP has a problem. But if that one server is the only HANA server in SAP, that’s catastrophic. You know, it’s this realm of contextual knowledge. Historically MSPs have that contextual knowledge, but it’s all the way at tier three and tier four. That contextual knowledge has to move to tier one. If MSPs want to get to the arena where that is no longer a problem, the contextual pieces have to move downward. You have to go from a hero-based MSP to a process-driven MSP. So many MSPs are built on heroes. It’s really hard to build a scalable business off heroes. You have to have heroes. Heroes are the people that when everything breaks and the world is on fire, they’re the ones who carry you through. And those heroes we want to have, we want to empower them, but they can’t be doing the stuff that should be done at tier one. So if we take that exact same question that you had, Rob, that question is, you know, how do we make, at the end of the day, how do we make MSPs more relevant to their clients and much more aligned with what the client’s trying to do? And that’s by taking the contextual knowledge of what the customer is trying to do, aligning that with the tactical approaches that the MSP is trying to do, and having a very crystal clear playbook of how this tactical component makes up this strategic initiative inside of the business. So we’ll take that, we’ll take that simple example. I shouldn’t say simple. SAP is far from simple. But the reality is, is that SAP is something that customers rely on. And when they rely on that, if SAP goes down the business goes down. And if you have an MSP that’s monitoring that, and at the same second of the same day, the MSP gets 36 tickets. We’ll just pick a random 36 number. 36 severity one tickets come in at that point in time. One of those severity one tickets is for SAP HANA. And the customer only has one instance of that. And that is taking down a large company. So that’s the first ticket. The next 35 tickets are for ephemeral workloads that the customer migrated off of, you got the alert, they migrated to a brand new ephemeral workload. And the 35 don’t matter. They’re false positives. But the one fully matters. In every single MSP on the planet, those 36 tickets are eligible for the same response interval. That’s a pretty tough average to be able to. Are you going to luck out and get the one? Or are you going to luck out, or not luck out, for lack of a better term, and work 35 false positives before you get to the one that matters? Now, most MSPs are going to tell me and they’re going to tell us that, well, we have more than one tier one path. That’s great. But the reality is you need to be responding to that one ticket right now. And you need to understand that that one ticket matters. And the only way you can do that is by starting at the beginning, starting with the sales cycle, understanding what customers are doing. If you’ve already gone down the path and the customer’s embedded, use your customer support teams. Understand what your customers are doing, start layering in that context, start enriching that data, knowing what that actually feeds, and understanding the dependencies and interdependencies inside of that. So if that server goes down, certainly you could by virtue say a database server going down is a SEV-1, but it may not be. If they have four database servers, they’re running in a high availability group, who cares? If one goes down, not the end of the world, go fix it tomorrow. That’s where context, that’s where understanding those dependencies is so crucial. And I mentioned at the beginning of this is how do you take that first step forward? We always take this first step forward and how I instruct MSPs is start doing things like this, take this step forward, break this down into simple programmatic approaches. And when we think about AI, it’s the exact same idea. We move steps forward, we have agentic, we have generative. Pick one, pick an area you want to focus on with your customers, understand the business outcome they’re trying to do. And if you have an inference engine, that’s going to be really crucially important here. So let’s understand that. Let’s monitor that. Let’s understand the intricacies related to how that customer is leveraging it, why it’s important. Are there latency constraints? Are there packet loss constraints? Those types of things. Let’s monitor to that and let’s understand how that happens. And if a customer has an application on the back end, you know, maybe they have New Relic or they have AppDynamics or they have some type of APM toolset, great. Let’s start bringing those into our monitoring. Let’s start bringing that intelligence in, understanding application flows, understanding dependencies, building that to be part of our story. And now we create so much more opportunity for us as an MSP driving that contribution margin northbound.

Robert Dutt: So it sounds like we’re kind of defining good visibility in a modern environment and kind of setting up for looking forward as understanding what actually matters to the customer and understanding what kind of flows into it, what all results in that thing that’s important to the customer still being up, still being running, still being functional, and kind of work backwards from there as opposed to the more “this machine is working, this machine is not” kind of approach.

Jeff Collins: Yep. Yeah. You want to go from tactical to transformational. That’s really the idea.

Robert Dutt: And you shared kind of the idea of the first step to do towards that. I guess as you’re moving towards that first step, you know, is there any one question or kind of mindset that you find works for MSPs to have in mind or asking customers to surface those blind spots and really start to understand what that context is that they have to have?

Jeff Collins: Yeah, that’s a really good question, Rob. And, you know, there’s some things that I do tell MSPs to start with before you ever ask that first question. One of them is kind of some of the simple, let’s call it research that you can do before you ever reach out to your customer. One of the easiest things you can do is start by what industry are they in. You know, in Canada, Canada has a lot of oil and gas, lots and lots of oil and gas companies exist in Canada. And so if you have an oil and gas company, we can start right off the bat with a lot of the things that oil and gas companies live and die with. And we’ll just pick on this one as an example. So oil and gas companies have SCADA networks. They have industrial IoT devices that are out there. They’re processing massive amounts of data. That data may be going into the cloud. It may be going into a data center. It may be going into some type of vault or something like that, depending on what they have. But each one of those are things that, as an MSP, you can start out before you ever ask your customer anything. You know that those are the things that exist in their environment. And you can quickly look and see, well, am I monitoring any of those? Well, no, I’m only monitoring Active Directory. Okay, Active Directory is probably important to the oil and gas company. But if it goes down, do they quit producing oil? The answer is probably no. And so if your answer is ever no, you know right off the bat that you’re not monitoring something that’s strategic to your customer. And so the first thing that you should always think about is, okay, if we have this industry, we should be monitoring the things that are strategic. Well, how do we do that? Well, we start with that one step forward. The first thing we talk to them about is just like when we went out and we sold that initial monitoring of Active Directory, they did it because they didn’t have time for it. There’s no oil and gas company on the planet that has time to be monitoring their SCADA networks. They just don’t. They may tell you that they do, but they don’t. So leverage your relationships, leverage your engagement with them and go after those pieces. Understand, you know, if they’re in AWS IoT Core, understand what that looks like. Understand who’s monitoring that. Understand how DevOps is working within that space. Maybe it’s DevSecOps inside of that environment. Understand that convergence of the teams and then start building a story around, you know, let’s take that on for you. Let’s start changing that. Let’s use the same paradigm that we’ve done, driving MTTR down, driving availability up, driving resolution times down, all those types of things. Let’s bring that into the era of SCADA networks, IoT, our core infrastructure. That’s where we start changing the value inside of our customer engagements. And that’s really where I see a huge opportunity for MSPs across Canada, where you can take that environment, you can take those opportunities you already have, and you can grow them from, you know, maybe you bill that customer $1,000 a month. You can grow it to billing them $20,000 or $30,000 a month, but it’s the most crucial $30,000 they spend. Because, you know, if that offshore environment or that, you know, oil sands environment or whatever it might be within the oil and gas space or in the energy sector, whatever it might be, those things are crucial to their business. And so the more that MSPs can kind of make that step forward, and then also start incorporating AI, every single one of those entities is incorporating AI. They’re incorporating it directly into their pipelines. They’re incorporating it directly into their data pipelines, not just the oil and gas pipelines, but each one of those, the more you can incorporate that, the more you can monitor, the more you can show value of everything that you do amazing as an MSP, that’s really where you start creating that intrinsic strategic value and you get out of that tactical approach.

Robert Dutt: And the good news is for a lot of these folks in the MSP space, presumably they have some of these pieces already in place, just not necessarily connected up to the technical side, i.e. sales and marketing have been focused on a vertical. And even if they haven’t, because they have customers in this space, they’ve built some of that muscle memory, some of that knowledge of what really matters. Now it’s just a matter, hopefully, of connecting it into the services that they’re offering.

Jeff Collins: Yep, totally agreed.

Robert Dutt: All right. Well, it’s been a really interesting look at sort of where visibility is at. And I think a real interesting opportunity that you’ve surfaced in terms of how it can be turned into a value conversation. I appreciate your taking the time.

Jeff Collins: Sounds great. Thanks so much for having me on, Rob.

Robert Dutt: There you have it, my chat with Jeff Collins from WanAware. I’d like to thank Jeff for sharing his insights. The thing that stuck with me from this conversation is how much of what’s changed in the modern network hasn’t been designed in, it’s been bolted on. AI workloads, hybrid architectures, IoT, SCADA, all of it layered into environments without the kind of formal rethinking that happened when we moved to virtualization. And Jeff made a really compelling case that for MSPs, closing that visibility gap isn’t just a risk management play, it’s a revenue opportunity, and potentially a significant one, especially in verticals like energy and critical infrastructure where visibility is tied directly to uptime, safety, and compliance. We’ll be back on Monday with In Case You Missed It, your weekly news roundup. Thanks for listening. I’m Robert Dutt for ChannelBuzz.ca, and I’ll see you in the channel.