.png)
MSP Mastery Podcast
Welcome to MSP Mastery Podcast, the podcast for IT leaders, MSP owners, and service delivery professionals who want to elevate performance, improve processes, and stay ahead in the fast-changing managed services landscape.
MSP Mastery Podcast
Delano Collins on Data Driven Decisions, Automation and Building a More Profitable MSP
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Welcome to MSP Mastery, the podcast for MSP owners and leaders who want to build a better MSP.
I am Jeni Clift, joined by my husband and long time business partner, Nick Clift. Together, we draw on our experience to explore what is really working in thriving MSPs.
In this episode, we welcome back Delano Collins, Chief Operating Officer at Electronic Office. Delano joins us for a wide ranging conversation about data driven decisions, automation, AI, profitability and the leadership needed to bring a team along with change.
Delano shares how his team has spent the past year improving processes and data integrity to better understand their business. He explains how combining information from MSP Bots, ConnectWise, financial reports and managers with Copilot has helped identify pricing issues, incorrectly allocated time and client agreements that need attention. Rather than simply collecting more data, the focus is on turning that information into clear actions.
The conversation looks at why sound processes must come before automation. Delano shares his approach to coaching people, explaining the why and helping team members become cultural torchbearers. He also talks openly about an automation that sent the wrong messages to clients, why the team stopped using it and what the experience revealed about training and communication.
We explore the hidden work that affects client profitability, including administration, account management and repeated touches on a ticket. Delano explains why he wants to understand what a healthy service delivery model looks like for his own business, rather than relying solely on industry averages. He also shares how clearer boundaries between support and project work can help protect both margins and client outcomes.
Delano describes how his team is using AI to analyse business information, develop a vCIO roadmapping tool and build a ticket triage agent. We discuss starting with manageable problems, choosing tools with a clear purpose and keeping human judgement, privacy and security in the picture. Throughout the conversation, the message is consistent: technology should reduce waste and improve outcomes, not simply add more noise.
Here is what we covered together:
✅ Why reliable data and sound processes must come before automation and AI
✅ Using visible dashboards to help teams understand performance and priorities
✅ Coaching people to understand the why behind accurate time entries
✅ Building a culture where team members take ownership and learn from one another
✅ Combining financial reports, operational data and business context with Copilot
✅ Identifying pricing issues, time allocation errors and leaks in client profitability
✅ Understanding the balance between billable and non billable work
✅ Why industry benchmarks do not tell the whole story about your business
✅ What went wrong when an automated ticket follow up process confused clients
✅ Using ticket age, activity and escalation to improve service desk accountability
✅ Setting clearer boundaries between support requests and project work
✅ Catching incorrectly coded project time before it reaches billing
✅ Using technology to reduce waste and support stronger margins
✅ Developing AI supported ticket triage and vCIO roadmapping tools
✅ Starting with practical automation opportunities while retaining human judgement and protecting client data
We created this podcast to share real conversations and hard won lessons from people who understand the challenges of the MSP industry, and to offer ideas that can help you build a business that is both profitable and fulfilling.
Follow Delano on LinkedIn: Delano Collins
Read more episode notes at mspmastery.blog
Watch on YouTube at youtube.com/MSPMastery
Listen to the audio podcast at youtube.com/MSPMasteryAudio
I believe positive people believe you're here for a reason and you're around the other people that you're around for a reason. Sometimes it might be to develop patience, but a lot of times it's to learn skill sets from people who are different from you. It's not practice makes perfect, it's perfect practice makes perfect. What is waste? Waste is anything that a client will not pay for. And so we need to use technology in that place to make sure that we have the lowest possible prices and the highest possible margins.
SPEAKER_00Welcome to MSP Mastery, the podcast for MSP owners and leaders who want to build a better MSP. I'm Jenny Clift, and alongside my longtime business and life partner Nick, we draw on our almost 60 years of combined experience to explore what's really working in thriving MSPs. Our goal with this podcast is to share the real stories and hard-won lessons that inspire and add genuine value to our industry, helping you build a business that is both profitable and fulfilling. This is MSP Mastery. Here's Nick, myself, and today's returning special guest, backed by Popular Demand, and currently sitting at our second most popular episode. Delano Collins is the Chief Operating Officer at Electronic Office with almost 25 years of leadership experience across managed services, cybersecurity operations and AI, is dedicated to helping businesses align technology and strategy to deliver genuine business outcomes. Having worked both within MSPs and with tech vendors, Delano brings a unique perspective that combines operational experience with a deep understanding of the wider technology ecosystem. Delano, welcome back to MSP Mastery. Please share your personal and professional best from the last six months.
SPEAKER_01I've built a platform for VCIO road mapping. We were talking about through the use of AI and two friends from the MSP space. That's been a labor of love and part, you know, in on us as a side project. But at Electronic Office, you know, we adopted Roost as an, you know, first we adopted MSP bots, and and that was primarily to see the data, to see how well we do what we do. And that ushered in an era of, I'll call it a renaissance, because we needed to understand where we needed to improve and understand, you know, margins and and utilization and who's bringing the most as far as effort to our company and which of our clients are, you know, where we need to bring the most effort to clients and what we need to fix and where we're springing a leak in the engine, so to speak. And that effort continues today. In fact, yesterday I spent a lot of time taking financial data, compounding it with information that I know. I actually took three different spreadsheets as well as contextual information that I know, as well as contextual information for my managers, adding it to Copilot. I built a spreadsheet with and had co-pilot build action items. It took me about 45 minutes to do all this, and it basically identified 17 clients with very specific action items on. On this one, we need to look at why we're writing off time. On this one, we need to look at maybe it was actually three clients why we need to look at why we have so much labor. On these, we have pricing issues. We have embedded engineers and we didn't price those correctly. On these, we have products and we discounted the products too much. And it was very specific information that we were able to provide because of the decision that we made a year ago to go with MSP bots. And it's a combination of financial data that we're pulling directly out of ConnectWise, the default financial reports, because of the way we've set up our agreements to be very transparent, and that information comes straight out of ConnectWise. And so it's part of this renaissance. And now I'm assigning account advisor. I need you to go look at this. Our service manager, I need you to go look at this. You know, CEO and I are looking at these other things, and we're we're tackling these issues. Spent two hours meeting with a client yesterday to talk about being data driven. They want us to come in and simplify their network. And I said, that's not that's not an A to B line that we can draw. The first thing we're gonna do is we're gonna be data driven and we're going to create tickets and then we're gonna present to you the issues that you have. We're going to we're gonna fix those things first and present to you the issues that you have. Basically, they've been a very tough client, but at the end of a two-hour meeting, which felt very tough, they were very complimentary on how we've been data driven, how we've been process driven. And all of that starts with the data.
SPEAKER_02Yeah, that's awesome, man. That's a good win is to get that data goal because that's what I the guys I coach and work with. It's about convincing them you need to be making data-based decisions, not emotional, not what you think someone needs or what the tech tells you somebody needs. You need to have the actual data to back it up. So yeah, good win.
SPEAKER_01The sexy words are automation and AI, and everybody chases those things. But if you don't have sound data, which are built on sound processes, then the the automation is garbage. The AI is garbage. Those outputs are always garbage. And so you've got to go back to the ground game and make sure that those processes are excellent. I mean, if you're a sports fan, you have to practice extremely well. It's not practice makes perfect, it's perfect practice makes perfect. And so you've got to make sure that when you are coaching your team, you've got to coach them and make sure that they go through the rhythms in the right way every single time. And you spot those things that could lead to errors in the data and make sure that they understand you've got to do it this way instead of that way because this could create a problem in the data. And then you show them the data and say, like, when you do it the right way, this is what we get. When I think of like what excites me the most, we've been transparent about that for a year. And now we've created torch bearers. Other people are you can see the the flame going off, the light bulb going off, and now they get it, and now they are cultural torch bearers. I spoke earlier before the the podcast started when we talked about people in our company who are using AI and they're getting it and they're running forward at light speed. We have myself, I use AI. I have an older Knock gentleman who is very adept at AI, but we have a 24-year-old who is a wizard at AI, and it's the data that has lit these people on fire, and they they get it now. And I don't have to babysit them. The manager doesn't have to babysit them. We give them a task and we're we say, go do it, and something that we think would take a week, they're 48 hours later, they're like, done, and then they demonstrate it, and we're like, whoa, we cannot believe this. And that they're going to our biggest client, and our biggest client's like, we've never seen anything like this before. That when I talk about with that, that as a CIO, I sit back and I think, wow, giving a task to someone and they come back with something that wows me, I there's never a day that that that doesn't just make me excited to go to work and see that kind of stuff.
SPEAKER_00And in the scheme of things, a year to make that transition, it's not very long.
SPEAKER_01You know, it's not, you know, the mistake I made earlier in my career was like, I thought that I had to like put everybody on my back and I'm like, I'm gonna, I'm gonna drag everybody ahead. I'm the technical guy, and I'm gonna, but this is what we're gonna do. And like culture suffers in that. But when you put culture first, when you explain the why, when you focus on your leadership skills, when you force people to do it themselves, and they have to learn the skills, and you focus yourself on. I've got to develop the communication skills, I've got to develop the patience, I've got to be kind and show grace and let them grow and make them do it. And we can't move any faster than the slowest person. And I've got to make sure that I put the right people in the right seat, and I've got to make sure I have the difficult conversations and make sure that the team moves at the pace that the team can grow. When when everybody starts getting it and everybody starts moving faster, everybody can move mountains faster and at a greater pace than a single person ever can. You know, teams win championships, individuals win awards, and and that's something that I tell my team now. And I think we're at the at the point where our team is ready to win championships. Everybody's playing at a much higher level than an individual ever can.
SPEAKER_02Wow, that's that's amazing. And I couldn't agree with you more. It's and it's that everyone moves at a different pace. And once you've given the go, get the people that are leaning in, you give them permission to go and excel. Other people see that and they get enthusiastic and think, oh, actually, I can try this too. And it doesn't matter if it I'm not as good as that guy or it doesn't work for me. At least I'm on the same journey and you share those with the wins. And like you say, the the teamwork is the teamwork makes the dream work, that's for sure.
SPEAKER_01Right, right. I believe positive people believe you're here for a reason and you're around the other people that you're around for a reason. Sometimes it might be to develop patience and communication skills, but other but but a lot of times it's to learn skill sets from people who are different from you. And, you know, we work in very diverse cultures and diverse work environments often, and that's a huge opportunity to learn from people who are very different from you. I know that as I when I started out from an engineer, my skills were very focused on Windows server environments, but I didn't know anything about SBS servers, I didn't know anything about firewalls. And there was a lot of things that I didn't know anything about Citrix. And I had very specific people who I can name right now who took me under their wing and were like, let me show you how to do this, let me show you how to do that. And I learned so much. I was a sponge and I was determined as a newlywed that I'm I have to learn these things. I need to learn these things. And this is a once-in-a-lifetime opportunity. And working in the MSP environment, I just just like a dry sponge that was given water, I just grew and grew and grew. And and I know some people don't like the MSP environment, but I I loved it. And and it was a it was a free education. I learned more than that I ever learned in college or tech school, which I speak incredibly highly of tech schools today.
SPEAKER_02So 100%. Okay. That's cool.
SPEAKER_00Okay. Quick personal win, and then let's get into our questions.
SPEAKER_01Personal win. Well, I live here in Western North Carolina and tucked up here in the mountains, and I've got three sons, and we do a lot of work to our house. You can kind of see behind me over here and around me. We've got a lot of deer and a a friendly bear. He leaves me alone and I leave him alone.
SPEAKER_00Is that a thing? A friendly bear?
SPEAKER_01You can go outside any day and hear roughly 22 different birds and stuff. But we spent the summer building paths, you know, flagstone pass. We moved probably 12 tons of material to build paths on the back and all around the side. And anyway, so we we did a lot of work this summer.
SPEAKER_02Well now the bear's got a path to follow. He doesn't have to go through the through the trees, you know?
SPEAKER_01Nothing seemed to stop him before. We we have a path to run away.
SPEAKER_02Go in this direction. I love it. I love it.
SPEAKER_00Okay. Let's jump into MSP bots. So when you made the decision to go down that path, what was the problem you were trying to solve? And the next part of this is how did how did you decide that automation was the right approach or that MSP bots was the right approach?
SPEAKER_01Sure. So when I first got to EO Electronic Office, we were not using dashboards and data transparency was not really part of the culture. I mean, they would they were emailing out automated reports to everybody and their brother from ConnectWise, but from my experience, emails are not are not actionable, and we get overwhelmed with too many emails. And so I wanted to move to TVs mounted on the walls and constant dashboards with numbers that are green, yellow, and red, like a stoplight. And if it's green, it's good, and if it's red, it's bad. And I needed to dial in on key metrics based on industry standard metrics, based on like kill rate and SLAs, and I wanted aggressive SLAs. I didn't I didn't want soft SLAs. I wanted to put myself in this in the shoes of a client. Like how fast are we resolving tickets? Like based on on criticality or priority. And there's some there's some other metrics that we look at, but like the age of the ticket, like aging tickets, like you know, and and how how many tickets do we have based on those age? What what is our utilization by engineering? Those are things that our managers and our leadership look at. But some we started off with just a few, and then we've added some over time based on that evolution. And then we had people who did not have good time hygiene. MSP Bots allows me to basically send somebody a message through team saying, like, hey, clean up your timesheet. Hey, enter your time in real time. And then we've extended that to the sales team that says, hey, you have an opportunity without a next step. Hey, you have an opportunity without a dollar amount. Hey, you you missed your next step. Those things have allowed us to like quietly prod and poke people into the next phase, but it started with the data. And that's really where I needed to start. I needed to start showing people what what I felt was, hey, here's where we are as an organization. This is good, this is not good. And once we were there, we were going to start like prodding people. I I knew I couldn't hire enough people to like go tell people, hey, this is what we need to do. We'd previously used other dashboards, dashboarding tools, but honestly, nothing felt as modern or current and integrated with Teams as well as MSP bots. And that was the deciding factor over why I decided MSP bots.
SPEAKER_02I think we used MSP bots too back at DWM. And then later on we moved to Bright Gauge. That's the same thing. You've got to have quality data to give to people. And I loved what you're saying, Gelano, about explaining the why of the data. And once we got that through once I got that through my thick head and I was actually able to explain it to the text, they changed their behavior. But you just can't just tell somebody to do something. You need to show them the impact of what their behavior or their activities is giving to the client or to the business. And once you've explained it out properly and shown them the data, I think they easily easily follow, follow the correct path. Make it easy to do the right thing.
SPEAKER_00This is what we were talking about before we started. As I said, I should have hit the record button. And we're talking about the the data that you're getting now. So can you share that with us again, that example of when MSP bots created that significant improvement for your team and the data that you're getting from that now?
SPEAKER_01So we started looking at the margin of clients, and we knew there were some issues there with some of our clients that were that didn't have the right margin, but we didn't know exactly what the issue was. And what we ended up doing was we took data from MSP bots, we took data from our our financial reporting, we took data from ConnectWise, we took contextual data from our managers, data, contextual data that I added, added that to Copilot, combined all of that, and then I basically told Copilot to build me new spreadsheets. And then from those spreadsheets, I analyzed those and then refined my commands in Copilot to basically not only build new new spreadsheets, but then interpret that data into actionable data. Then I went and verified that data and built a priority list of what we should do. That's been one avenue. Another avenue has been interpretation of pricing, interpretation of pricing specifically for managed services, different managed service packages, managed services for embedded engineers, how much how much time we should be spending on our account advisors on different clients of different tiers. So our biggest clients, what that ratio is. What I was working on yesterday is I've been looking at billable time as versus non-billable time and what that ratio should be. There is a ratio of how much time it takes to support a client. You have non-billable time that you support clients with, and you have billable time that you take, but there's an unhealthy ratio. And we saw in just a real-world example, the most unhealthy clients that we we have, there is a 2.73x ratio. One one is, you know, the one is billable. The 2.73x is unbillable time. That ratio is wildly unhealthy. Most of our clients are well below a one to one ratio. And so what I was talking with the leadership team, demonstrating through the data that we're getting out of out of not, it's not just MSP bots, which is all of this is feeding in through AI. We're basically pouring this through AI and then building new new spreadsheets. I'm able to demonstrate this through the clean data that we it's really taking us a year to get to. I'm able to demonstrate that we have there's a ratio that we are zeroing in on that is a healthy ratio of billable to non-billable time. And that anytime, once we zero in on this ratio, we need to flag any client that gets above this ratio. And we need to understand why. We spent a lot of time yesterday talking about do we need to tweak how we how different teams code their time so then we can understand if a project person codes their time this way, or if the administration codes their time a different way, or if the account advisors code their time a different way, how do we flag that ratio? What should that ratio be for each person and our team relative to a healthy organization? So you you see how we're getting deep in the weeds, and we want to understand how much salt goes in the soup to make a healthy soup. We we need to understand what is our formula for delivery for service, and and and this is what's coming out of all of the data in the review that we are we're consuming right now, and this is part of the renaissance.
SPEAKER_02Yeah, that's awesome. And and I think you're right, because I have always had access to lots of industry averages, and you know, historically it's mostly been US through TruePeer and Evolve and Comp Tier and all those organizations, but there's some of it's regionally based now, but uh that gives you a benchmark or a or a kind of a target to go to if you don't know anything about yourself. But what you're doing is much more powerful because the industry average ratio doesn't understand exactly your client mix, your team mix, your geography. So the work you're doing is gonna make give you a competitive advantage because you're gonna know exactly what good looks like for you and your contracts and not have to compare it to the guy next door or the guy down the road because he could be working in a completely different model. So yeah, that's really interesting. And it'd be it'd be awesome at some point to catch up and figure out you know where did you land on that? Is it because in my head it's 80% billable, but then again, yeah, you've got account management, VCIO, sales admin, concierge, you know, inviting people to the football or whatever, all that work happens for a client that you charge a submanaged service fee, but that work should not all that work should be charged to the contract.
SPEAKER_01I worry a little bit about like industry standard times because there's a whitewashing that occurs there. Because I know in some MSPs that I work, I've worked with, worked for, they don't some people don't bill their time. They just come in and they're like, I'm gonna charge code eight hours, boom, and they walk away. And and so there's a whole there's teams of that that contribute to this non-billable time that never actually charge time to the client. We don't do that. We make everybody bill their time. I don't care if you're ordering a workstation for client X, like bill your time to that. So we're tracking it all. It it is a little bit of uh extra overhead, but we have incredible data integrity in what we're doing. And so I feel like I'm in a unique, you know, position to see exactly how much admin, you know, which is a huge bucket. I almost hate to use that term. Customer service, customer service, yeah. Yeah, and with the with the recent just skyrocketing of pricing, like we will someone will won want to buy a workstation and then they there's a delay in payment, and then we have to go back to the vendor and we're having to tell them, oh, there was a 13% price increase. And they're like, What? You know, hold on a second, I need to go get approved. And so by the time they get approval but have it paid, there's been like a 5% increase. And so there all that delay means the person and admin who Was ordering that workstation spent like they touched that ticket five times. And so there's like four hours in trying to order a workstation. And but we track that time. And so I go back and look and I'm like, oh my gosh, like, like, how do we change our processes to adapt to this insane, like, skyrocketing pricing market, you know? And I look at some of the inefficiencies with what's going on in this crazy pricing market. So that that's this is when it people are like, oh, well, you're the the industry says that your ratio should be. I'm like, I don't care what the ratio I've been in this industry too long to care about what people tell me things should be because I I know from a sanity standpoint, you know, when what people tell me, oh, my margins are blah, blah, blah. I'm like, I know, I know how all the song and dance goes with like, you know, getting your company listed in the top 500. I know, I know exactly how all of this stuff happens. So I'm not too worried about that. I'm more worried about running a good business that delivers great service and makes employees happy at the same time. And that's hard enough.
SPEAKER_00I keep everybody happy. So let's jump over the other side of the fence. So automation doesn't always go to plan. Can you tell us about a project that failed either to deliver the result that you expected, or perhaps a client project? And rather than specifics that what went wrong, how did your team respond? And what did you learn?
SPEAKER_01The the most obvious one is there is a there's a status that we use, an automation status, that when a ticket is, when a client doesn't respond to a ticket for a certain period of time, it will start sending notifications like, hey, your ticket's important to us, but we haven't heard from you. And it'll do that for three days, and after the fourth day, it closes the ticket. And what what I noticed when I first got to EO was we had a lot of tickets that clients were not responding to. And I'm like, I'm gonna solve this right now, I'm gonna implement automation, and all we got to do is I'm I'm gonna create a new status, and I'm going to set these automations up. And on the on the fourth day, after three notifications, it'll close the status, it close the ticket. Well, what happened was some of the engineers were not using the status correctly, and it started firing off to clients who should not have been in a waiting on client status. And then next thing you know, we're getting these hot rocket emails from from clients who are mad, like, what do you mean you're waiting on me? I'm waiting on you. And you know, the service manager is like running in my office, like, what is this thing doing? Like, you know, uh, and so then uh we're I'm going, I'm spending time in his office, like, hold on, let's what's going on? You know, well, Bill put this ticket in the status. I'm like, why is it in the status? We're looking at the ticket notes. We're like, we're not waiting on the client, they're waiting on us. We got to go talk to Bill. Bill, you're not supposed to be using the status. You know, we sent out the email. Bill didn't read the email. All right, everybody stop. We ended up we tried to space it out a little bit more, but you know, engineers are engineers and they don't always read the fine print. And, you know, we still couldn't get people on the same page, and so we had to stop using that status. And I still love the idea of that status because clients are sometimes really hard to get, and and we have tickets that just get stuck. And what do you do with those? But we have we have a dispatcher, and she calls the clients and calls the clients and calls the clients, and eventually she will make the decision on whether to pull the plug on it. But you know, I I like the idea, but if it's not a perfect world, and so we had to back that one off.
SPEAKER_02Yeah, we we implemented a very similar similar thing to that in cons consultation with our clients, and then we had some pushback from a couple of hospitals that had lots of part-time staff. So they said, well, three days is not enough for us because some of these people work three days on, three days off. Can you extend it out and make it yeah, every second day you send a reminder? So it's a seven-day and that worked well for us as well. Like my draconian method is if it's a service test ticket, it should not be over seven days old. No excuse for a service test ticket. If we've got a client that can't work for seven days, we have a big problem. And if it's oh, well, I fixed the problem, they're working, but we need to go do some other work on something else. Well, then that's a different ticket. That's not the same ticket.
SPEAKER_01I love what you said there, Nick. You know, people have asked me how many tickets should be on the service desk board? And I said, that's the wrong question. I said, how old should any ticket be on the service desk board? That's the right question. And when when you ask that question and you go look at the age of tickets and you look at tickets older than a certain period, like I don't think tickets should be older than 10 days. If any ticket older than 10 days, for instance, if you if if I asked you to fix my printer and it's 10 days old, like I'm already mad. Like there's a problem. Any ticket that I've asked you to fix that's older than 10 days, that tells me something about your service. And there's a problem right there. And when I go look at tickets that are older than 10 days, those typically tell me there's a there's a deeper issue that we're not able to solve. And that's when all of those need escalation. And what I found, the honest truth, we have clients who reject projects, and then they try to send those tickets in through the support desk and get our our support engineers to go implement what should have been a project. What I have, our dispatcher coaching our our service desk manager and coaching our engineers is saying, like it's not the engineer's job to say no, but it is your job to like raise your hand and go to the manager and say, I don't think this should be a support issue. Support's job is to fix something that was previously working. And if it's not, if it was never previously working, stop and go talk to your manager. And if the manager doesn't feel like empowered to say or have a conversation with the client, go escalate it to the next person up the chain. And I'm more than happy to have a conversation with the client. And I've had several conversations with clients saying, respectfully, like this is not covered in your agreement. And here's why. We have a team that's specifically designed to deliver great outcomes around resolving this. It does cost money because they're highly trained around those outcomes. But the support desk will not deliver the outcome that you want, even though you think it's going to be free. It'll cost you more money in the long run with downtime and delays, and you know, a lot of like pulled-out hair. Like, you don't want a person who's going to touch this and then go away, and then three days later try to touch it again, and then next week try to touch it again. It's going to be the biggest nightmare. It'd be like a haircut that like spans three three weeks. Like they're going to cut a little bit and like come back. Mate, exactly, you're exactly right.
SPEAKER_02And I I've had this exactly the same conversations with clients. I said, Yeah, we could do it on the support, but there's no guarantee, and it's going to be piecemeal. And that's a great acknowledgement, a haircut over three weeks, you know. You really want a haircut done in 15 minutes so you can get the result and enjoy it, not not wait three weeks for the whole damn thing. And the the other while we're talking about numbers, the other thing that made a big difference for us was we didn't worry about we went away from worrying about the age of the ticket so much and we went to tickets with no activity. So if you have a stale ticket that has no time entry, no updates, no notes for more than I think it was 16 business hours, so two days, that the service coordinator or the manager, the team leader, was talking to the engineer saying, What is going on with this ticket? And if you do that constantly, automatically the old tickets get resolved because humans want to help people, right? So a new ticket comes in, they want to do the new ticket, they don't want to go back and look at the old ticket. I already patched that one up. Mary's happy. I'll just leave that ticket and one day I'll get back to it. And that one day never comes. So we put that loop in where there's no activity on the tickets, and that was a metric on a dashboard that we looked at on a daily basis, and that made a huge difference. The old tickets just kind of went away. I know MSPs out there that have two, three-year-old tickets on their service desk. Like it's ridiculous. Like, oh, but one day we need to fix that. Yeah, but one day's not today. That's not the service desk's job. That's an account manager of ECIO. That's a sales guy's job to have a conversation with the client. It's not we're not the dumping ground. Don't let your service desk be the dumping ground, guys. That's what happens a lot.
SPEAKER_00So let's get back into MSP bots. What's the most surprising outcome that you've had in your service delivery? So we talked a lot about data and and what you've seen coming through and making data-driven decisions. But yeah, a really surprising outcome of an maybe an unexpected benefit, change in the way that your team working, something along those lines.
SPEAKER_01Well, I mean, we've definitely caught some problems that we never knew we had, where we had people on projects who were trying to do the the right thing for clients, where they felt like we were billing too much on a project. So they went for a managed service client. So they went in and they actually coded some of their time to the managed service agreement under the project, and it was flying under the radar. And what we were able to do with the MSP bots was create an alert that says anytime there's a project that has any time coded to an agreement, it immediately notifies that person that says, you're not allowed to build time on an agreement under a project, change it now. And then if it doesn't, if they don't fix it within two days, it it notifies the person and the manager. And if it's not fixed within two days after that, it notifies me and basically the administrative manager who also runs HR. And then we have a chance to fix it before it it hits billing. And so there's been a lot of things like that where we're catching things that we never knew were leaks in the system, you know, in the engine where we're just, you know, leaking oil. There's so many aspects to, you know, uh ultimately our job is to manage resources and make sure that, you know, energy gets converted to revenue, and we've got to make sure that all these little leaks get addressed. And MSP bots, I think, is an exceptional tool at making sure that I don't have to pay a person to go tell people to do stuff. If I see it, I can just create a bot or an action that tells person in real time, like, hey, don't do that. Fix it, you know? And it's correcting activity in real time. And I really like how that works. And so that's been an exceptional. But, you know, I think the real story is MSP bots is just step one. Step one really allowed us to take step two, which was taking the next step with another platform roost. And then from there, we built, we we focused on clean data with all of our agents, integrations with gradient, and automated billing. So we trust the numbers that go into our agreements. And so we don't have someone like making sure that, oh, this client has 53 agents and they have 53 EDR, and they have a backup with, you know, X and blah, blah, blah. And so we're we we are automating those counts because we trust those counts because we've done our due diligence there. All of this stuff is part of an ongoing effort to trust automation, to trust the supporting data. And we are moving forward with being a real uh company that uses automation, that uses data, that uses technology to drive our business to be more efficient. We have to drink the Kool-Aid of technology. And where where do we need to do that? We need to do it every single place that that drives operational efficiency. I think back about what is waste? Waste is anything that a client will not pay for. And so we need to use technology in that place to make sure that we have the lowest possible prices and the highest possible margins. It's not just about the tools, it's about the right tools driving the right outcomes to create the largest possible margins at the lowest possible prices.
SPEAKER_00Whatever business should be doing, isn't it?
SPEAKER_02Ultimately, that's what we as MSPs want to be doing for our clients is helping them use the technology that we support and provide to improve their businesses and their profitability and their clients' experience. So yeah, there's a I was listening to I've got a bit of a summary of there's a a guy from I think it was a Caseya-sponsored AI presentation that was run around the traps last week from a guy called Dan, some long name in America. I can't remember his surname, but it was really interesting. And and he's saying the same thing. It's like AI is a pattern matching data collection tool. It does not make decisions, it does not invent things, it's not creative, but it can process all the information and look for patterns. And and the way you use it, you need to be cognizant of how it operates. It's just learning, you know, GPT stands for generative pre-trained transformer, something right about that. So it can only produce data that it's been trained on. So it can't invent any new uh information. It's only it can only do pattern matching. Having said that, that little ramble. Has there been any processes or tools that you have been able to improve or replace by using some form of uh agencai yourself in your own business processes, like using Copilot or Claude to process ticket triages or match up reports, etc.
SPEAKER_01Yeah. So we looked at an industry tool. We were going to buy something off the shelf to triage tickets, and then they basically changed the requirements of the tool because we use on-prem ConnectWise Manage, and they basically said you can't use this with you've got to move ConnectWise Manage to the cloud. And we have some other requirements that we would have to change. That looked like that was gonna be a very long endeavor to move ConnectWise Manage to the cloud. Not that we don't want to do it, it was just gonna take just the movement of the data of managed to the cloud was gonna take us a long time. And so we thought we're not ready for that. And we're we're already adopting Roost. We, you know, we're and we're making more progress with clients in that regard. And so what we wanted to do, we've had so much growth with AI, we thought we could we could just adopt Claud, our own enterprise license, and build it ourselves, honestly. And what we were seeing with Roost and the new release of Roost, we we thought there's no reason we can't build this ourselves. And so we're we're building our own triaging agent ourselves, and we're gonna take the Roost portal and actually drop it in our own website and do what, you know, use some of the Roost workflows and have our own triaging ourselves. So there's there's that piece. We're currently working through that now. We're also building our own VCI road mapping tool, and we're using Copilot now to assess profitability and look at specific patterns within agreements, like I mentioned earlier. And that's been very helpful. Up until this year, I had a very poor opinion of Copilot, but they've made some tremendous strides there. And I'd say that you know, what I've been able to do with Copilot in the last three months has been tremendous for our business. If I could, I would show you what we've been able to do there. What I've done in Claude is I've built industry calculators to show you like what you can say by converting, you know, outsourcing versus your IT to our services, and we're using that on our new website. And so I'm basically building stuff for our web designers, and our new website's gonna launch gosh, any second now. So instead of like having a conversation and telling a web designer, I'm able to go build it and then give them the code, and then they basically add the polish to it.
SPEAKER_02So yeah, the prototyping and the proof of concept is just unbelievable. Yeah.
SPEAKER_01I remember years ago sitting around a table with marketing people and trying to tell them what I knew, and then they would go create the stuff, and I'm like, no, no, no, no, I didn't say that. Now I go to Claude and I build it, and I'm like, and I I basically add all the context to it, and then I give it to a marketing person, and they're like, they just add this or just tweak that, and they're like, oh no, if you just add this color combination, it'll look so much better. And I'm like, great, you know, and so I can get it 95% of the way there, and they add their 5%, and then boom, we're ready to go. And so going back to what we talked about before, the the research shows that there's a lot of assumptions about what AI can do. If you don't know anything about what you're looking at, AI is incredibly helpful. It helps you move roughly 19% faster than what you would know. If you know the answer if and you use AI, it actually slows you down. But in cases where you have a lot of data and you need to assess a lot of data, I find it extremely helpful. So, like in an executive who's empowered to use AI, we've had incredible results. And you know, when I need to combine three spreadsheets and I need to like get very specific information from those, within 15 minutes, I've got what I need. I've got a new spreadsheet built with the exact columns, with the exact correlation. And then I'm walking into a meeting. I can actually tell it to build a spreadsheet and then build a corresponding PowerPoint. Usually the PowerPoints aren't exactly what I need, but I can go fix those very quickly. This I remember this used to take me days to build this stuff, and I can be ready for a meeting in in under two hours now with everything I need. And it that this is part of how we're moving so fast because I've got over two decades of experience. I know what I need. I just need to build it and I need to show people the data, and the data is what's convert what's changing opinions and and helping people get over their biases about the business.
SPEAKER_02Yeah, exactly. Right. And and you mentioned um earlier that you would you'd built a VCIO like road mapping tool. Is that in test? Can people check it out? Is it I'd I'm myself personally, I'd like to have a look at it.
SPEAKER_01Sure. VCIO Blueprint.com. So we're we'd love to have people check it out.
SPEAKER_00Okay. So last question for an MSP owner who is, I guess, considering with automation, how do they make sure that they're genuinely improving service delivery, for example, if that's their problem and still supporting their team, delivering a great client experience, not just getting more noise. So where do they start?
SPEAKER_01I'd say you've got to make sure you pick the right platform. You've got to make sure that your team is entering the right data, and you've got to make sure you've got the right team members who can help you, you know, plant the right pivot foot. A lot of the right platforms will help guide your way forward, but if you don't have the right team members, you're not gonna get started on the right foot with automation. But I'd I'd say start with the start with the biggest pain points, but if there's simple pain points, you know. If there's something like restarting systems every night, restart those systems. If it if it's something like freeing disk space that owns systems, free disk space. You know, you've got to start somewhere, start s somewhere simple. When I think about some of the complex things that we wrote, the the complex things were little pieces of things that we it took us years to learn how to build. And we had to like daisy chain those more complex scripts.
SPEAKER_02Yeah, and I remember like talking about a mistake when I learned about an RMM and we signed up for an RMM and I gosh, it was a long, long time ago. So I I know what you're gonna say.
SPEAKER_00Still have a nervous twitch about it.
SPEAKER_02I think it was enable and central and was installed locally, and I said, This is amazing. Look at all this data. So I turned it all on and blah blah blah. And the next morning there was 2,000 tickets in this for this desk.
SPEAKER_00Because when he said he turned it all on, he means he turned it all on.
SPEAKER_02Yeah, I turned it all on. So yeah. So we were getting like alerts for everything, and there was no alerts for alerts, yeah. And then I had a quick call with the guys and said, Ah, I think you should put it in learn mode first, not just put it in alert mode. So yeah, I made that mistake and everyone freaked out. Like a core service manager came into the office and there was 2,000 tickets. And we we back in those days we were probably getting 200 a week. So yeah. Yeah, you've got to start with the end in mind and queue off little bits and and make improvements over time, not just go the big bang theory. Yeah.
SPEAKER_01I remember when we first signed up for LabTech, which has become automate, and we talked to some other MSPs, and they all uniformly said, Oh, it'll take you a year before you really get any traction on using this. I think we all got a little nervous on did we make the right decision on moving forward with this?
SPEAKER_02But we did, and we doubled down on it, and you know, that was uh For the guys out there listening, or the people, girls and guys, I would say pick a tool for solve any problem that's worth solving and just embrace it and get into it. And I see too many people sitting on the fence and they get three months into a product and a new serenity object comes out. I was a hundred percent guilty of this, and you get distracted, and instead of implementing the first one properly, you already kind of mentally decided I'm gonna change this and go somewhere. Somewhere else. Then you go down that path. And by the time you get three months into that one, another new shiny object jumps out over here. And sometimes the devil you know is better than the one you don't. And yeah, the old grass is greener on the other side of the fence. Just double down on what you do. Do your research, do your due diligence. But once you pick a product, just get into it. Build a good relationship with the vendor. Don't be adversarial. Like there it's in their interest to help you make their product good and frame the conversation that way. So hey mate, I'm I love your product, but it's not delivering what we need. You need to help me sell my service manager on your product again because it's it's got these three problems. We're trying we don't understand why it's not doing this. Work with them. And and if you ultimately can't solve the problem, then today you can vibe code something or you can find another AI tool or use your fellow MSPs. And you know, there's lots of different chat groups out there, lots of everyone's got their own personal signal and WhatsApp groups with five or six MSPs in there that show us share very intimate details about what works and what doesn't work. And yeah, and there's every week I learn about two new tools that are like, oh my god, not another show any tool.
SPEAKER_01Yeah. When it comes to trying to determine what to attack, I'd say there's two approaches. The the database approach is look at what is your big what are your biggest problems and what of those can you solve? And it and then the other is, you know, what are the easiest problems to solve and go solve the quickest, easiest problems. And there's, I mean, uh, there's different schools of thought, but if you can, if you can solve 10 problems today that are just quick little easy fixes, you'll see an ROI from that eventually. Or you can go try to uh tackle that one big mountain today, and it might take you two weeks, and you'll see a huge improvement once you tackle that that big problem. But you know, you may not have the skill set to tackle that big problem. So just consider that. You know, what can you tackle? You know, either the smallest things that you can knock out right away or the the biggest thing that you need to knock out right away. So but either way, those both of those approaches have a work for me. Sometimes I'll give the big project to the most talented person, and then the other people, I'll like distribute the small task, and everybody's you know, working towards the middle.
SPEAKER_02Yep, they're all contributing, yeah, 100%. I think it's just human nature to there's not too many people out there that just do the work day in, day out and don't think about, oh, there's got to be a better way of doing this. And we need to find those people that think that way and support them and encourage them to be creative and obviously taking into account security and privacy and all that kind of stuff, but be creative in how you solve the problems. Using AI is a great tool to help you solve problems, but I wouldn't go uploading the client's payroll data spreadsheet into chat public chat GPT and asking how how can I get a pay rise, you know. That's not that's not maybe not the right way to do it. But there are tools and don't don't be scared to use them. But as an MSP, having a policy is important. If you don't have one, people are using it anyway. And you know, like finding out who's using what AI tools, you know, the whole shadow IT thing, shadow AI thing is is an interesting topic that we haven't really touched on today. But understanding that your your your team and your clients' teams are already putting data into AI. So if they're not educated and they're not understanding what they're doing, they could be putting, they could be compromising data out there. You never know. We just found out today, wasn't it? Didn't OpenI in the US hack Medicare or something? It just came out, or maybe that was in Australia. Some something some big medical system got hacked by one of the AIs and it took them three months to admit they did it.
SPEAKER_01There's been two breaches by OpenAI, and one, it did take them months to to admit, although I I don't have the specifics in front of me.
SPEAKER_02That's our job as MSPs to educate our clients and our team on what the dangers and the benefits are. I'll remember the first days of cybersecurity, AI is no different. It's just it's just another tool to use in your business, and you've got to use it properly.
SPEAKER_00Because we run under EOS Nick's rock for this quarter was to automate five problem areas in our business. And you know, there's just three of us in our business, we're very small now. But what it did was made us because every week on our level 10 meeting, Nick was saying, What do you want me to do? What do you want me to do? Because I've got this rock to to automate five things, and I've got about another three weeks left in our quarter. And you are working on the fifth one now, Nick?
SPEAKER_02Uh about the seventh one, actually. It's really cool because I mean the great thing about using an AI tool to vibe code a uh a solution internally, like building a SaaS product and selling it to externally and all the security that goes with that. I'm not doing that, but internally, this is just an internal tool to help us manage our basically a mini PSA CRM, but it integrates to a whole lot of different bits and pieces, and it's just so cool. Like they say, Oh, I want uh we'd be better if you could do that. And within five minutes, I've I've fixed it, I've tested it, and I've rolled it out and documented what I've done. And that would take months, weeks and months of work in the we wouldn't even do it, not for the size of our business. One screen on one platform. It's crazy how much stuff you can get out of it. It's also crazy how much time you can waste down rabbit holes at things that don't do anything.
SPEAKER_00Yes, you are very, very good at going down rabbit holes, Nick. But I think but my point was that having Nick having that rock and then pressuring by Goose, who's our marketing person and myself to be thinking of things that we can improve and automate and you know and the podcast is one of them because you know, there's a process of you know, who do who do we want to have as honest guests, like reaching out to them, booking them, recording the editing process, and then going through the you know the marketing of those and then the release. It's a process, and having that, we've now built a uh an app to manage that process. Where before we were using one of the off-the-shelf project management tools, which was really, really clunky. So you know, we've got some really nice little tools now just through Nick having that rock and constantly reminding us. So there's there's ways to, I guess, encourage the team to be thinking of things that to improve in your own business.
SPEAKER_01Well, you know, it's funny how AI has become just a part of my everyday habits. You know, one of my older sons, I've got three boys that he's an avid gamer and he uses a third-party plug-in. And the third-party plug-in does a lot of things, but one of the things that it does is inventory all of his game assets, and he's got a lot of different characters. And it even though it inventories it, he's got like thousands of things in his inventory, and he was like wondering, like, how do I even like even though it inventories it, how do I use it? And I was like, Well, let me just Google this, and I was like, oh, well, your inventory tool builds all the data in a JSON format. I was like, hold on a second. I was like, I was like, let me open Chat GPT. I was like, I browse to the the JSON data. I was like, I open, uh, I loaded all the JSON data and it I was like, and I said, inventory this data, and I was like, what are you looking for? And he was like, well, I'd like it to tell me like what I should use for this and this and this. I was like, all right, boom, boom, boom, boom, boom. It goes through, and then within within 10 minutes, it built a table of everything he needed. And so I was like, Chat GPT can is now an extension of any application in any an area of your life. You just have to understand its capabilities. And so even if there's not a native integration, it can read raw data and interpret that data, and it's incredible what it can do.
SPEAKER_02And it's it's also incredibly stupid, too, because this demo this guy did yesterday, he said if you ask any AI model, you type in the words once it you type in once upon a, what's it gonna come back with? It's gonna come back with time, right? Because it's trained on and then but the word after time, it's it's just a it has no idea what's gonna come next. It'll go and and you know, apparently 90% of the time it's once upon a time, I is the most popular answer. So it it's it can't invent a story, it can only regurgitate what it's been told. And that's what that that was the takeaway I got from that discussion today, is you've got to really think this this is not this is just pattern matching, it's just regurgitating what it's been trained on, it's not creative, it's not inventing anything. You need the human in the loop to, like you said, get the data, present the data to help with that for sure, but you've ultimately got to look at that data and make a decision on what to do next.
SPEAKER_01I I do worry about the cognitive impacts of AI because I I do feel like it makes people less likely to read, less likely to spend time absorbing information. It makes us less creative because we just want to go to AI and like kick it out. And so what is ultimately where is this leading us as humans, as creators, because the real value for humans is sweat equity. And if we if we're too lazy to put in the sweat equity, then where does that take us? I think about my personal journey. I got where I got because I had to work a lot of jobs that I didn't want to work, and I had to work a lot of hours that I I didn't want to put in. And and it was the banging my head against problems that I didn't know how to fix, and I had to try this and that didn't work, and I had to try that and that didn't work. I had I remember reading, remember those Cisco books when I went through Cisco Cat, I had to read all the Cisco books. I read, I read the 1100 page how to build and repair your own PC book. Like, who's gonna read that in the age of AI? Like, if I didn't read those books, I would have I don't know that I would have gotten where I got in my career. But all of those experiences and all of that education that allows me to sit down and pro and and have the perspectives I have. I'm not saying all my perspectives are correct, but I I I bring a unique set of perspectives, education, experience and history, yeah. Yeah.
SPEAKER_02And I agree, and I worry for the 24-year-old guy, like we got these experienced people that have the knowledge now, but like in programming, for example, no one's actually writing boring code anymore. So you have your high-level senior architects that design it, and then you just offload the code into an agent. But so you'd no one's employing software developers as interns anymore because they don't need them, right? The AI agents can write code just as good, if not better, than an intern agent. But when all the architects die, how how are we growing the new ones? Because the AI is not going to invent the problem to solve. It's going to solve a problem, but it's not going to it's not going to see the problem to solve the problem.
SPEAKER_01That brings to mind two points. There they used to say the the time to the time to detect vulnerability in any software used to be like, let's say 100 and something days, 180 days, and now it's negative 10 days because of AI. And I think, but I think it's not just AI, I think it's like the poor coding practices because people are in such a hurry to rush code out and you can vibe code stuff. And I think the vibe coding is done by people who conceptually understand, but they don't they they don't have those the secure coding background, and so they're jamming stuff out. There's a gentleman who has a a very large presence on YouTube and he's a a security expert. He's been incredibly critical of Fortinet, not to specifically call them out, but he's but he said Fortinet used to be a very secure ecosystem, but they've been plagued by a lot of insecure code. And he's been he been he's been very outspoken on YouTube about what's going on at Fortinet. Like, why are they building by default insecure settings in their firewalls? Like, why do they keep having all these issues? If you and he said if you look at the root cause of their issues, like the code is by by default insecure. And and if they were using the same practices that they were that they were doing 10 years ago, like these would they would not be having the same issues that they have. And I wonder if this is indicative of the same vibe coding, you know, rush to publish software practices that we're seeing throughout the industry. And this is why I think two of the last Patch Tuesdays had over 600 patches. Last month had the largest Patch Tuesday in the history of Patch Tuesdays. And so we're struggling to get clients to turn on their computers because they don't understand. Like we've got almost 700 patches this week to put on your system. They don't understand it's not like, oh, typical Patch Tuesday, we've got 23 patches. No, 686 uh 687 patches this week alone are coming down the pipe from Microsoft. And it's following a week where there were roughly four. Time to move to Linux. So it's unreal what's happening. And who's talking about this? I I don't know of anybody who's really talking about the content is AI sloped, and people just stop reading it, and no one's paying attention to really what's going on.
SPEAKER_02So the good news is the three of us are real live people.
SPEAKER_00We are, we are not AI. Delano, thank you so much. It was great to uh to catch up with you again and really appreciate you sharing your time and and expertise with us.
SPEAKER_02Always a pleasure.
SPEAKER_00And Nick, thank you.
SPEAKER_02No worries, yeah. It's been a bit of fun. Good to do one at night, actually. Yeah.
SPEAKER_00Yeah, yeah. We're uh 12 hours apart, so uh yeah, and Delano's having his morning coffee and Nick's having a wine, so refill. If this conversation hit home for you or got you thinking, head to mspmastery.blog and keep the conversation going. You'll find all our episodes there and more wisdom from the peers and partners who are shaping the future of our industry. And make sure you subscribe so you don't miss future episodes. We've got plenty more great guests and stories coming your way. Until next time, this is MSP Mastery.