5K Technical Services CEO Kirkendoll On Becoming Customers’ ‘Chief AI Strategist’

‘One of the things we’ve been really pushing out there is becoming that chief AI strategist … just like a virtual CISO, we’re that virtual person that is out there coming in to assess, making sure you get all your ducks in a row. And then as you go down that path, really handholding you, walking you through the process,’ 5K CEO Corey Kirkendoll tells CRN.


Generating marketing posts, screening resumes and helping nonprofits more quickly apply for grants are some of the early artificial intelligence use cases 5K Technical Services CEO Corey Kirkendoll is working on with customers. Down the road, AI improving video-powered security and video-content retrieval for customers will be key use cases as well.

The CEO of 5K—a Plano, Texas-based Microsoft partner and member of CRN’s 2024 MSP 500—told CRN that solution providers need to adopt AI practices to evolve their business and become better trusted advisers for customers.

“We have become this chief AI strategist for them,” he said. “We don’t know it all, but let’s sit down and have some conversations of where it comes from. … A lot of these guys, they just want to be better than their competition.”

[RELATED: Microsoft Earnings: CEO Nadella Says Systems Integrators’ AI Work Is Sign Of A ‘Maturing’ Market]

5K All In On Copilot

Copilot is still in its earliest days for Microsoft solution providers—Copilot for Microsoft 365 became generally available for enterprise customers Nov. 1, and Cloud Solution Provider program members have only had the ability to sell Copilot for M365 since January.

During Microsoft’s latest quarterly earnings report, Chairman and CEO Satya Nadella shared plenty of data points to show that Microsoft AI continues to grow in adoption, including a nearly 50 percent increase in the number of Copilot-assisted interactions per user in Teams and 30,000 organizations that have used Copilot Studio to customize Copilot for MM365 or build their own, up 175 percent quarter over quarter.

But Nadella also said that AI has helped fuel other parts of the vendor’s business, with cloud offering Azure now the “port of call for pretty much anybody who is doing any AI project.”

That’s sort of been a significant help for us in terms of acquiring even new customers, he said.

Microsoft has more than 400,000 partners, including systems integrators, resellers, MSPs and other partner types, worldwide, according to the vendor.

For Kirkendoll, some of the ripple effects of AI conversations with customers have been more customer interest in his data classification and security offerings. Most of 5K’s customers are in Texas.

“The precursors to AI are still a cybersecurity awareness and a defense state that makes sure that you have a good, healthy posture before you move forward,” he said. “And they’re not mutually exclusive. They’re together.”

He praised the amount of training resources Microsoft has provided its partners related to Copilot. But the MSP CEO still sees room for improvement in easing the cost of experimenting with Copilots.

“Those who have the operational maturity levels to understand that they need to have that to be better will pay for it. No problems and no-brainer,” he said. “But some of the guys still need to be able to have access to the product at no cost, not their bottom line. Because at the end of the day, Microsoft will benefit from us selling it and doing it, but they’re also going to benefit from us because we’re going to use it and become experts on it.”

Here’s more of what Kirkendoll had to say to CRN about the early days of Microsoft Copilots and the channel.

What do you think of Microsoft’s various Copilot business models

consumption-based Copilot for Security, per-user M365 Copilot, role-based Copilot for Sales, for example?

The consumption model is going to be one that I have to grow into. That’s not something that we normally are used to in our industry. … The per-user model is easier for us to bake in and to go down that path.

We’re leaning more toward that. … But at the end of the day, it’s really going to come down to not really focusing on the price, where it comes back is really continuing to push hard on what we do from value selling and why you’re doing this.

The price becomes a moot point. If I’m trying to sell you Copilot, I’m going to talk about what it does. … That $30 uplift on top of all the rest of my licenses can get really expensive from that perspective.

It is causing us to go back to the drawing board and really look at the whole of how we’re pricing this out. Coming out with Business Premium … make sure we have the right [Microsoft Entra ID] P2 license. … We are in that dangerous spot from where we come into what I call that ‘nickel-and-diming’ kind of a situation. ‘You want that, that’s $30 more. Oh, you want that, that’s $5 more.’ … Our customers don’t like that. … I’m trying to get to a point of pushing everybody to a better standard of license … and just letting them understand the price per value that they're going to get by adding certain things.

Does it have to be everybody? Can it be just these users? … That makes it tougher for us because not everybody’s going to have to have it.

I have got users who have Basic licenses because all they need is access to this. But then I have got users that are more into Standard, more into Premium, more into E3 licenses.

That’s the hardest part for me right now—really figuring out where these guys fit so that, one, I’m giving them value. But I’m also not draining the bank with just licensing. That’s the hard part—finding that sweet spot of what that looks like for us.

How are customers reacting to Copilots and AI so far?

We’’e seeing everybody come out with a pilot program and we turn these five users or these three users up, which is what we’re recommending. … But even before we do that, we need to start having a conversation on policies and procedures and what that looks like.

The main conversation we're having right now is the data classification pieces, which is pushing us up to a DLP [data loss prevention] conversation.

Even with that small pilot group, ‘Hey, let’s make sure that you’re good because even that pilot group may have your executive assistant, someone in HR, you as the CEO and somebody else.’

Now you’ve opened up Pandora’s box where you all can see and look at stuff that you probably don’t want to have access to. Even that little, small pilot group can become dangerous.

So it’s forcing us to do the DLP side, forcing us to look at Business Premium right off the bat and really focus on how we lock it down, even in that small pilot group.

For the most part, we’ve had some CEOs just turn it on … just to play with it, get a look at it and do some things.

Is this a good mix of revenue opportunity and finally persuading a customer to improve their security?

One hundred percent. … I believe that AI and these conversations are going to drag the security conversations forward.

So I still believe that the precursors to AI are still a cybersecurity awareness and a defense state that makes sure that you have a good, healthy posture before you move forward. And they’re not mutually exclusive. They’re together.

And by doing that together, you can move down those paths and figure out where it needs to go. But … it causes the customer to start to get their offers and their workplace in order to clean up the junk.

It has been just laying there, hanging out there … willy-nilly. Now when you start bringing in these structures and start really using the data, start analyzing the data, you go, ‘Uh oh. Maybe I didn’t want that to be transferred out. Maybe not everybody should have had access to that.’

Because remember, a lot of these customers, they set this stuff up even before the MSPs came in because they thought they could do it. Turn on Office 365 and away it went.

So now we’re trying to unravel that and we’re caught in the middle of saying, ‘Hey, you remember when you set that up?’ Because you bought an Office 365 license, turned on OneDrive and SharePoint and you were happy, ‘Oh I don’t need an IT guy.’

And then all of a sudden, now we’re in there, we’re like, ‘Hold up. What’s going on here?’ We have got to go back and try to fix that. So trying to protect them from that perspective is probably one of our biggest challenges.

And for us, that's 100 percent project engagement to do this. We’re not going to get rich off Microsoft licenses by any means. It’s the project work and the managed services work that gets on top of that is what makes it [worth it] for us.

Has Microsoft shown that it understands the importance of solution providers in meeting this AI moment?

The way Microsoft has come out with … Copilot guides and training and things, this has been probably the best place I’ve ever seen them come out and really get ahead of this and provide you with true end-user friendly as well as partner-friendly materials to go out here and do this the right way.

It's almost like ‘paint by numbers.’ If you pay this, pay this, pay this, you’re going to get there. … They’ve really put the training wheels on it to help you get out and be successful.

But it really takes that MSP to be diligent, to follow the path and not think that they’re smarter than they are and get out there and get in trouble.

And then leaving it with the customers, what I’d love to see is that they need to engage with a professional as they do this. It is very easy. But you need somebody that’s experienced to make sure that you don’t tlet the genie out of the bottle.

You are ahead compared to other partners when it comes to adopting and piloting Copilot

are you normally an early adopter of technology while other partners and customers wait to see it proven out?

This is why I go to a lot of shows … I love to figure out what’s coming next. I worked in big companies. I understand what disruption means and … always be on the forefront.

So I did that with security. Doing that with AI. Did that with IoT. … Figure out a way to really bring that value back to show that we are truly that trusted adviser, not just the PC guy.

I truly believe that we evolve. It won’t be that point of doing support when, ‘Hey, my email’s not working.’ … We’re just going to be the guys guiding and helping them deliver and stay on track to where they’re going.

And a lot of this is going to be how do we differentiate them in marketing? How do we differentiate them in an HR perspective? … This is not a technical conversation. This is a business problem conversation. … I’m spending a lot of time with the individual business owners talking about, ‘Tell me three to five problems that you have and we can figure out how Copilot and/or some AI can help you solve that.’

And now that we've solved that, now you have an HR issue where you have got to figure out how do I repurpose these resources to go do something different?

And then making sure that I’m keeping them abreast of all the laws and all the other stuff that’s coming along. … I’m working in front of lawyers like, ‘Hey, you have to make sure—it is trust but verify—that the AI is not making up stuff.’

Those are pieces that we’re trying to do. They don’t have anybody that’s doing that through governance. We have become this chief AI strategist for them. We don’t know it all, but let’s sit down and have some conversations of where it comes from. … A lot of these guys, they just want to be better than their competition.

They want to be doing what their competition is doing. They don’t want to be first. They definitely don’t want to be last. But I have got to stay in front of them so they are aware of it.

What are some cool Copilot use cases so far?

Internally right now we work with a partner where we’ve been creating LLMs [large language models] for MSPs where we can specifically leverage … how we handle responding to cases and getting things back and notes and creating documentation and SOPs [standard operating procedures] for ourselves, which now brings time back for my engineers—’Hey, you feed it here and document this or make this email better or document this SOP based off of a framework that we fed it before.’

And it creates that, saves the documentation and keeps moving on. Because I know as an engineer, we don’t like to document. … We support a lot of nonprofits, one of the challenges we’re having right now is looking for grants, qualifying for grants and taking that grant-writing down to a point where it’s more manageable.

And so that’s where we see a huge case for Copilot in Microsoft 365—all of those grants and other documents, all of that information, is stored into your SharePoint site or your OneDrive site.

So now I can come up with some situations that allow Copilot to reach into all of that documentation and pull out statistics and pull out information that will not be readily available.

Today, literally, they have got books sitting on a shelf, paper that they have got to go back and say, ‘I remember that grant from a year ago. I remember that from a meeting I went to at that time.’ … We're taking the process of writing their grants from six to eight months down to maybe two to three weeks. … We’re in that assessing stage of understanding what their true problem on the grant-writing looks like.

And so we’re writing out, one, what the prompts and what they look like, how they fit and where they offer up from that perspective.

Then we’re in the process of categorizing the data based on the feedback that they have for what they need to research to fulfill the grants. … Now it’s also at a place where it’s going out there to find the grants that it knows that I qualify for and helping me write those. … We’re probably just past the beginning, in that proofing out stage right now.

It has been really cool from that perspective—just to see the amount of data that they had that they didn’t even know they had, stuck in emails and PowerPoints. … [The internal AI] is in that beta stage. … It’ll probably be in full-blown use probably by May, June. But it’s in the beta stage, testing it out, tweaking it, working it into the workflows. … There are eight [full-time employees]. … We’re working with our marketing people to show how we leverage AI. But we’re also showing how we’re doing our due diligence. So we become our own case study to our customers. … ‘Here are real-life scenarios on how we’re using it’.

What’s 5K Technical Services’ strategy for starting and growing AI projects with customers?

We go out to marketing—CMOs and directors of marketing—and show them how they can leverage and create posts. … And then when we get a chance to do it, we spend some time talking to the HR side. How they can screen resumes and look for resumes, create job descriptions and things from that perspective.

And then we mosey around … we’re looking for specific business issues that they’re seeing where they have got a human or somebody has taken reports or is analyzing data and it’s taking a long time to get through.

Once we are able to develop and automate that for them, the light comes on and they say, ‘OK, what’s the next one? What’s the next one?’ … We’re taking that physical resource, turning it into digital labor. Now they can take that physical labor and put it on something that’s more meaningful. It’s almost like a two-for-one. I’m gaining [a resource].

I tell them all the time, the guys who are hiding, they are going to get lost. There’s nothing I can do to help them. The ones who are smart are going to get smarter. Those who are already [AI] smart, you’re going to become geniuses. … That’s what AI is going to give them.

What types of customers are jumping into AI?

Most of these guys are our long-time customers. … We’ve been talking about pushing people to the cloud, people still want to hang onto the servers in the server room. When I’m able to do this, I need to push that data into the cloud so we can have access to it.

It [AI] is driving those projects and conversations that have been stuck and moving those forward.

This is one of those vehicles that continue to push forward all of the things that we’ve been asking for. … It is not just coming in with security and they are like, ‘Yeah, I get it.’

But now they have got a need. Because if I do this, now I can tie some ROI dollars and [they can] say, ‘This is directly affecting my bottom line, my resources and getting more stuff done.’ … I just happened to be dragging security along, kind of that Trojan horse behind it, that Jedi mind trick. … What Microsoft has done uniquely is—everybody’s heard about the open stuff out there, ChatGPT … But they don’t control it. … What we talk about is controlling your data.

It’s already back behind your infrastructure. And what they don’t really know is because they’re in the Microsoft Graph today, it’s happening. It’s just waiting to be exposed. You just haven’t done it yet.

So when you’re talking about how easy it is, it makes that transition and to be able to adopt it that much easier. And what’s cool is that people who did go through our hardening and go through those things, it’s even easier for us because it’s almost like we knew they were coming. So it makes it easier for them to adopt. It makes us look better. It makes us look smarter. It makes it look like we’re really watching out for them as they go forward.

But I will say the biggest thing for us is we are the first to have these conversations before somebody like me comes in and [the customer] says, ‘Why didn’t Corey come in and talk to me about this?’ … Dallas is a heavily saturated MSP market. Very much so. But it’s a wedge for us because now I’m having these conversations. Not necessarily saying, ‘Hey, you have got to have an MDR [managed detection and response service] or an EDR [endpoint detection and response service].’ And they go, ‘I don’t know what that is.’

Hey, you want Copilot? No problem. Let’s throw in Intune—now I’m not talking that gibberish. Now it’s like, ‘Oh yeah, no problem.’ … I’ve been trying to get you an EDR and Intune for the last three years. Copilot was your way in.

What are some future use cases you are excited for?

We do physical security, and we do a lot of audio-visual stuff for churches. And so one of the things that happens is the church is always looking for old footage and things to tie back to an anniversary or some celebration that happened.

We are leveraging AI to go back and look at data from a video perspective. … We’re also looking at tying it into our physical security to where AI is watching and counting who came into the sanctuary, who left the sanctuary from a security perspective. … If I can see what’s in your hand or see what’s out there … there’s a lot of things that I can do to alert [authorities] well ahead of time, whether that’s a weapon or gun or kidnapping or something going on.

Having AI look at it is crazy. I have got a church [client] that has probably 120 cameras across the campus. They put it on three 65-inch screens, but … it’s still that small.

And by the time they figure it out, it’s almost reactive. They say, ‘Hey, something happened on camera one, can you click into it?’ By the time I click into it, it’s already gone.

But if I had AI looking at that, AI that saw it in real time and was able to alert but then also go in and say, ‘I saw him on camera one, camera three, camera seven and camera 12’—and pull it together and give you one complete picture of what it was.

Very, very cool stuff versus me going and saying, ‘Oh, this was at 13:55 and this is at 13:57.’ By that time, the guy’s gone. It’s a nightmare. I have missed the whole situation. But now I have got AI saying, ‘Hey, I can correlate those images and give you a full timeframe from all those seven points of view just by asking me what that was.’ … Catalytic converters are getting stolen, so one of the things we’re seeing is, ‘Hey, if I have got that camera … as long as I have a feed from it, I can tell you if someone’s coming in.

I can see if the truck is moving, I can do all of that stuff because AI is looking at that to be able to alert and catch these guys. They’re quick. I think in all of about three and a half minutes they can rip a converter off and be gone. That’s about the time I have got. And then it’s like thousands of dollars to go fix it. It’s crazy. And they could get hit over and over again. … AI can save that cost all day.

Anything you want to see from Microsoft looking ahead?

There needs to be a partner NFR [not for resale] use case for this for better adoption. … Those who have the operational maturity levels to understand that they need to have that to be better will pay for it. No problems and no-brainer.

But some of the guys still need to be able to have access to the product at no cost, not their bottom line. Because at the end of the day, Microsoft will benefit from us selling it and doing it, but they’re also going to benefit from us because we’re going to use it and become experts on it.

That’s the challenge. That’s the key that I would love for Microsoft to see. I agree, sell it to the customer, we will figure out how to do it. But me as an MSP or service provider using it, give me a copy so that I can go become an expert.

This eliminates one of those barriers where I have got to sit down and say, ‘Hey, do I have to pay for this?’

Has your distributor been helpful with Copilot adoption?

Pax8 is a unique beast out there. … They have got some great workshops they have put together. They have got some great documentation. … They continue to help us strive to get better at it.

They've trained up their group to be ready to help us on certain things. Pax8 is awesome at that point.

Working with nonprofits, how has it been moving them onto NCE?

That question only gets hard for the MSPs that don’t spend time and talk to their customers.

Because we actively stay in front of them and let them understand and know that prices change … it has been an easy conversation going forward.

It’s different when you haven’t said anything, haven’t seen them or anything and now you have got to send them a bill … It puts you into a defensive mode that you don’t need to be into. Then you are not really being that trusted adviser that you need to be. You need to be bringing it to them ahead of time. I’ll send an email out [about] these new MFA [multifactor authentication] bombing [attacks] that you’re seeing on Apple phones.

I say stay in front of them because I don’t want somebody to call and say, ‘Hey, has your guy told you about this? Nope.’ And then now the door is open to have that conversation. ‘Well, he’s probably not telling you about something else.’ … We all know that it costs a fair amount of money to acquire a customer. It doesn’t cost much to continue to keep the customer and to sell more into. And that’s that loyalty that comes back.

Do you want to see the Copilot margin go higher?

I would love to see it go up, more points is always good at the end of the day. But for me, I just don’t want them to change the model or start adding things to where it changes my engagement and consulting model from what we do as project-based.

Because it’s a long-term engagement. We’re consulting and showing them [customers] how to drive that and go down that path. So I’m cool. … At least that’s a little bit worth going in having this conversation that it is not freely added on.

We’re already moving down that path, putting them into Premium, getting them set up right, getting their licenses right, and making sure they’re protected. … One of the things we’ve been really pushing out there is becoming that chief AI strategist from that perspective. … just like a virtual CISO we’re that virtual person that is out there coming in to assess, making sure you get all your ducks in a row. And then as you go down that path, really handholding you, walking you through the process.