Listen to the podcast
Read Transcript
Erick and Rich discuss N-able’s Shadow AI Visibility solution and the larger opportunity it represents for MSPs to add AI services, plus why and how to build a multi-quarter AI and security roadmap for your clients today. Then they’re joined by James Farrow, CEO of Cyft, for an insider’s take on the present and future of AI productivity tooling for MSPs. And finally, one last thing: why the absolute worst time for a pizza delivery robot to arrive at someone’s home is when a SWAT team is already there.
Discussed in this episode:
N-able Sheds Light on AI’s Shadows
AI agents are breaking things and organisations know it. They are deploying more anyway.
New Research Finds Workforce AI Policy Compliance Is an Enforcement Problem, Not an Awareness One
DoorDash Delivery Robot Invades SWAT Scene, Won’t Leave as Cops Flashbang Resident
Some guests on this podcast are clients of Channel Mastered. Compensation plays no part in their appearance or the content of the discussion unless the episode they appear on is a “bonus episode” explicitly labeled as sponsored.
Transcript:
Rich: [00:00:00] MSPs, summer is here, and so are vacation schedules, staffing gaps, and the challenge of keeping service levels high with fewer hands on deck. Join MSP industry thought leader Erick Simpson and Kaseya’s Dan Tomaszewski for a live discussion on how successful MSPs are leveraging AI and automation to streamline operations, reduce manual work, and support more clients without continuously adding headcount.
You’ll discover practical use cases for AI-powered service desks, ticket triage, security operations, and faster issue resolution that can help your team work smarter while maintaining an exceptional client experience. Save your seat today at https://bit.ly/aimspwebinar. That’s [00:01:00] https://bit.ly/aimspwebinar and learn how to run lean, stay profitable, and scale support this summer And now, on to the show.
And three, two, one, ball. Blastoff. Ladies and gentlemen, welcome to another episode of the MSP Chat Podcast, your weekly visit with two talking heads talking with
you about the services, strategies, and success tips you need to make it big in managed services. My name is Rich Freeman. I am Chief Analyst at Channel Mastered, the organization responsible for this show.
I am joined, as I am every week, by your other co-host, who also happens to be our CEO and Chief Strategist at Channel Mastered. His name is Erick Simpson. Erick, how are you?
Erick: I’m doing well, Rich. How are you? And you’re celebrating a win recently for your ball club. They beat the Angels the other day, and yeah, that series is going on strong.
Rich: I, the so that… the game you’re referring to was Monday. They I [00:02:00] believe the Mariners won last night as well. Shh. But the Monday game was special, folks, because for various reasons, Erick’s wife and his two sons and their girlfriend, they were all going to that game, too. And I just happened to have tickets, and so we got together for a beer beforehand and and then I got to watch my home team win, and they got to watch their home team lose.
So the evening ended better for me than it did for them. But it was a very nice evening just the same.
Erick: Yeah. Beautiful. And what surprised me was I was getting pictures sent and saying, “Look, it’s 8:00 PM right here.” And it was, like, bright daylight, and here it was, Southern California’s complete darkness, right?
So- Interesting … that’s pretty wild.
Rich: Yeah. Yeah. Obviously we’re further north here in Seattle. I’ve always wondered how different, how much difference does that make, and there you go. That striking, ’cause yeah, it was still daylight when I left that game.
Erick: Yeah. We’re we’re thick-skinned, w- we Angels fans. We’ve had to build that resilience of, not being to a championship in [00:03:00] over two decades. Not… we won the World Series, I think it was in 1980 or something. It was that long ago. Anyway hope still burns strong in true Angels fans.
I was also surprised at how many Angels fans there were at the game, ’cause, I was watching it at home, and there were tons of Angel fans dressed up and representing at at the Mariners stadium.
Rich: I go to about eight games a season, and yeah, it, pretty much every game I go to, there is good representation from from out-of-towners, and that’s partly, because Seattle is a tourist destination.
So I’m gonna be in town anyway, I’ll go watch a baseball game.
Erick: Absolutely. All righty. Well- Now let’s get down to business, Rich. Let’s- It’s time to kick off this session. Yeah …
Rich: let’s get into our story of the week, Erick. And this comes out of a conversation I had last week. Regular listeners, watchers know I was at Ruth’s conference in Nashville last week.
While I was there, I met with the CEO and president of N-able, Jon [00:04:00] Paluca, and their chief innovation officer, Robert Johnston. And we spoke about a new release from N-able called their Shadow AI visibility tool. Didn’t get a lot of fanfare for some reason or other in the media. But before I talk about what it is and why they’ve introduced it, and why they think it is differentiated from the competition let me just give you a few numbers.
You know how I like my numbers, Erick. Researchers from The Economist recently published some data. 98%, 98% of organizations have experienced a disruptive agent-related incident of some kind already in the last year. Two-thirds of the organizations The Economist polled lack full visibility into their AI agents.
So having trouble to some scale, don’t even really know how many agents they have, what they’re doing, who’s running them, et cetera. Here’s some new data from AvePoint just came out within the last day or two as we’re recording this here. Approximately 20% of organizations they polled don’t know whether [00:05:00] employees are using unsanctioned AI tools or not, so they don’t know if there’s shadow AI happening within the organization or not.
Spoiler alert, there is. The number goes up. It’s 21.1% don’t feel like they know what’s happening. Are there shadow agents at work there? The number for gen AI is actually 17.6, a little lower than 20%, but that’s up from 6.3%. So organizations- are less confident about their visibility into what kind of shadow AI usage is happening within the organization.
And then last but not least, Erick, this is my favorite. This comes from a company called Neon Cyber. 63% of survey, and I’m quoting you this ’cause it needs to be understood verbatim to just soak up the goodness here. 63% of survey respondents reported having a clear AI policy that they understood.
However, of that 63%, nearly 50% knowingly violated it by using unapprived unapproved AI tools [00:06:00] anyway. And there’s no visibility at management, so management doesn’t know what I’m doing here. I– There’s a policy, I’ve read it, I understood it, and I’m ignoring it because I like this tool and it’s unapproved and management really can’t do anything about it ’cause they don’t know that I’m using this unapproved tool.
So that obviously is the setup for why MSPs need a tool like the Shadow AI visibility solution that N-able has rolled out Key word in there in that name is visibility. This is an observability tool right now. So it will go out and it will let you know it will inventory all the shadow AI tools that are in use in an end user organization right now.
For now, it’s not gonna do anything about that. So you as an MSP, you’re gonna have visibility into what’s out there, and then you’re gonna need to have a conversation with the customer about what to do a- about that. And N- N-able has a pretty strong point of view, basically. Even if we gave an MSP the ability to just shut down those [00:07:00] unapproved tools, they don’t actually think that’s a smart business policy for MSPs right now because the people using these tools are probably using them productively, so you don’t just kinda wanna shut it off.
What you wanna do is work with the customer to enable safe usage of these tools and train people who maybe think they can only get a job done on this unapproved model. So it becomes a consulting opportunity, and that’s an important part of the the message N-able is putting out there right now, which is if you are an MSP looking for a way in, a, a- an AI service that you can add to your service roster right now, a good starting point, we’ve talked about how many MSPs are looking for that, inventory shadow AI.
Ju- just like you used to do vulnerability scanning for current and potential customers before, now you’re gonna do shadow AI scanning. Go in talk to the customer about the best steps forward to enable responsible, safe use of AI inside the organization. That’s gonna produce project work, and then on [00:08:00] an ongoing basis, you’re gonna continuously monitor shadow AI usage, compliance with the policy et cetera.
Project revenue followed by recurring revenue. At some point down the road at Lumen, the NDR XDR solution N-able makes will have the ability to enforce the policies TBD on the timing for that. But for right now, it’s a visibility tool that creates an AI opportunity for MSPs. Now, obviously, Erick, we’re familiar with a number of different tools, standalone tools from startups that do similar things, except they do go beyond observability into enforcement and guardrails and and that kind of a thing.
N-able will tell you if you are using our RMM, they call them UEM unified endpoint management tools if you’re using our MDR XDR solution at Lumen, getting that all on one pane of glass native integration there efficiency benefits to the MSP in doing that. And obviously, that is a point of view shared by companies like ConnectWise and [00:09:00] Kaseya that have a f- fuller suite of products out there.
That’s the N-Able point of view. But I think for our audience today Erick the key message is the numbers I shared before make clear your customers really need help with this issue to a degree they don’t know. This is a, an AI- a value-added AI service that you can offer that will produce short-term and long-term revenue, and there are tools from your existing and potentially from new vendors out there that can help you do that
Erick: You had me, Rich, at 98% reported agent related disruption.
Can you unpack that a little more? Is that just like anything that has caused some sort of a… ‘Cause to me, a disruption is a business interruption. Like I’m stopping what I’m doing to go deal with this problem, whatever that problem is. That seems insane to me. Almost 100% of the folks, it, that were polled said, “Oh, yes, we’ve had agent related disruption.”
Rich: Yeah. And I, I don’t have a breakdown of the [00:10:00] specific kinds of disruption, and I don’t think, I don’t think they’re talking about the kind of disruption that shuts down the business. It’s the kind of disruption basically where the agent did something suboptimal, something that you didn’t instruct it to do.
May or may not be calamitous but it to put it in your words, it’s something that you have to stop what you were doing and deal with, even if it’s only for a few minutes.
Erick: And that’s why I framed it that way, because obviously it’s, if it was a, a business continuity that affects the business, we’d hear, we’d hear it a lot more about it.
But that in and of itself, when you compare it against what people are saying, “Oh it helps me improve my efficiency and productivity this much,” you gotta weigh those two things. It’s what’s the true cost here, as well as token burn and the other things that we’ve been talking about.
But yes, we have spoken about other solutions that are standalone. There’s opportunity, right? Where there’s a need, there’s opportunity. So we’re seeing this happening in the channel, where we’re seeing other folks coming in and saying, “We have a solution that not only will help you assess and identify, but then create these guardrails and help [00:11:00] control.”
What I really like about N-able’s position was that, that statement about, or that position regarding don’t just shut it off. Assess everything, and then to me, as a former MSP, that’s an opportunity for me to put on my VCAI- VCIO hat, my consultative hat, get that conversation in the boardroom and say, “Okay, look, here’s what we’re seeing across the, across the environment here.
These are all the tools, the shadow AI tools, and these are the the things that people are using. So let’s have a conversation about that and identify, should we implement a policy?” And I… That other component was, yeah, the majority of these folks, whatever the nu- the number was 60-some percent said that there is a policy, but I am violating it on purpose because I need to use my tools.
I made a comment I think on our last podcast where it said, I said I’m now in the camp of one of those people that says I cannot live without AI.” But, with great power comes respon- great responsibility, right? So having that conversation and then there’s an, there’s a policy, yet [00:12:00] there’s no accountability, there’s no ability to measure, hence there’s lack of- potential enforcement of a policy, right?
So this is a big conversation that can be had in a very strategic manner with clients to prove your value, moving into kind of that next phase of just a value that we’re delivering to our clients. We’re talking a lot about getting out of kind of the commodity conversations and things like that.
Of course, we handle, infrastructure management and security and, help desks for users. But, here’s something that we need to map that is much more strategic to your business that can align with your growth goals, and then build that into road mapping and ongoing consultative services.
And really the idea that we’ve been talking about a lot on the podcast, Rich, around monetizing AI, right? And, we’ve heard all kinds of different, ideas and ways to monetize it, per agent, per workflow, per outcome, you know- [00:13:00] On top of, some additional percentage on top of token usage and all these things.
But to me, that’s the more the transactional stuff. When we can say we’re going to… the real services attached that we want to apply on top of maybe the way we’re generally monetizing some of the work we’re doing in AI should be those additional consultative services like you mentioned.
Now I’m in the boardroom, now I’m being more strategic. Now for every dollar of, that I’m realizing from my base AI, management, auditing, assessing, and then identifying, I want six, seven, $8 worth of additional consultative services attached. So it’s that ongoing strategy building, and then interviewing, different business unit leaders and identifying what it is that they’re doing in their workflow.
And then for the folks that are using AI, sitting down and having conversations with them, “Hey, how are you using AI? What is it you’re trying to get out of it? Maybe we can help.” So being additive with kind of that security mindset [00:14:00] that says, “Okay, here’s what the, your users are trying to achieve with these AI tools.
Here’s a way that we can support it safely and securely so we don’t have, agent related disruption or data leakage and things like that.” So these are the things that I’m thinking about now.
Rich: Yeah, and I, you make a good and important point. We’re talking nominally about a tool, Shadow AI Visibility from N-able, but this isn’t and shouldn’t be a tool conversation.
It isn’t between us here, and it shouldn’t be between the MSP and their customer a tool conversation. It should be a strategic conversation, and you’re absolutely right. We, we reiterate this regularly that it’s more important than ever, s- specifically, especially in the context of AI, for MSPs to have consultative relationships with their customers, and this is an opportunity to have a consultative strategic conversation with the customer about an issue with real security implications inside the business.
And then y- you can guide [00:15:00] them as the vCaIO to a better, safer, more productive future. It is a revenue opportunity, it’s a stickiness opportunity, it’s an opportunity to shift the dialogue with the convers- the customer more in the direction of strategy away from the transactional and that is all good stuff.
Now speaking of good stuff, Erick I’m sure you are aware we are recording this here on July 1st, which means welcome to the second half of the year. And just naturally you start looking ahead to the future and nobody wants to be surprised by what the future holds, and I think that has something to do with your tip of the week.
Erick: Nice setup, Rich. Thank you. Yes and yes, we are in a new quarter now, right? So we are now in Q3 of 2026, and so for our MSP listeners out there, what is your strategy to make sure that every meeting that you’re having with your existing clients provides value [00:16:00] beyond kind of, “Here’s my report card, mom and dad, I got a B+,” right?
So we’re talking about, and I co- I continually wanna remind MSPs that, yes, strategic business reviews are the gateway to building and strengthening your credibility, your value, and helping guide your clients’ budgets and use of technology, and doing the things that you recommend that they do.
But it is difficult to have a quarterly business review with every single one of your clients. So I don’t wanna so much turn this into a QBR conversation as more of a technology and AI roadmapping conversation for MSPs, right? Because sometimes there is a service or solution that rises above and beyond, just waiting till we ha- our next QBR.
And you’re not gonna have QBRs with every one of your clients. I assume that many MSPs out there do what we ended up [00:17:00] doing when we were mature enough to really figure out how To meet the needs of our most important clients, and then our growing clients, and then the clients that just really didn’t have any kind of growth strategy or budget and things like that, the clients that were flat year over year, never consumed more services from us.
We did a strategic business review as needed with those folks. But those are the C customers in that pyramid of ABC I talk about, Rich. So the tippy top the sect- section, I pyramid into three sections, right? The tippy top, smallest little tiny pyramid on top is the, are the As. The fat middle is the Bs, and the big base at the bottom is typically Cs.
And what we’re trying to do is trying to, move more into B and A clients over time, right? And so the B clients, you’re maybe meeting with them, again, depending upon their growth strategy and their budget, maybe twice a year, maybe once a year, who knows? But your A clients, sometimes we met with our A clients, Rich, like weekly.
They were growing so fast, they were doing things, [00:18:00] so quickly. Some we were meeting with monthly, some quarterly, right? And then beyond that, there were the B clients and things like that. So if we’re talking about AI and we see this as a risk, because it is a big risk, especially, with this new reporting that you just shared with us, Rich.
What’s happening in our client environments may not wait for a QBR. So this is a meeting that we want to have maybe in between the next QBR. And set a meeting with clients and saying, “Hey, we’re helping our clients move into utilizing AI safely and effectively and productively, and we just wanna have a quick chat to understand what your position is and help identify w- what AI is already being used.”
And sometimes, Rich the s- the business leaders are the ones using AI, right? They’re using it, and it’s always, it’s also funny when we do these surveys and things like that, like who are the folks that kind of fall prey to some of [00:19:00] these phishing attacks and email threats, and it’s typically, I won’t say typically, but a lot of the times it’s the business leaders that are doing these kinds of things. Anyway comment below and let us know if that’s what you’re seeing as well, audience. So every client needs to know where your relationship is going with them, right? Because if they’re just, if you’re just it’s the same thing every quarter, it’s the same thing year over year- you’re putting yourself at risk in that relationship because somebody come- may come knocking on the door, as happens, and says, “Hey, what are you doing about AI?”
And now we know that folks like Anthropic and others are going to be, knocking on your clients’ doors about AI and things like that. So if you’re really building a s- a strategic partnership, a business partnership, rather than, your technology advisor or your trusted advisor clients want…
they trust about seven people, right? According to Jay McBain, there’s seven people that are in the [00:20:00] circle of trust and, your technology organization should be one of them, but you have to earn that trust, and you have to be in lockstep with what that business owner’s vision is for growth, and their and their strategy, and s- support it, and put the guardrails around it so you can help them achieve it in a safe and productive manner, leveraging technology.
It’s not a tool conversation like you said, Rich. It’s not a service conversation. It’s a strategy conversation. So every client should know what’s in their roadmaps, and you’re doing a 12-month roadmap. You’re doing the basic things, the license management, the, the the asset management, and you’re replacing equipment and things like that.
That’s expected. That’s like getting your oil changed in your car, but what are you doing to improve the performance of that vehicle? That’s what this conversation is about. So three things you can do today as you’re having conversations with clients. At leading with AI and security, we’re seeing a lot of partners, Rich, being much more successful and [00:21:00] finally getting the security solutions in place because their clients want AI.
So this is a conversation that says, “We can lead with AI, but we’re gonna pull through our security to support that AI conversation.” And when we’re having these meetings with clients, it’s all about prioritizing, identifying your top client accounts, identifying and scoring them. Who are the folks that are most apt to take a meeting with us that turns into our ability to do an AI assessment?
According to today’s conversation, let’s stay with that. An AI assessment, identify shadow IT, and then present a process a service offering that you’ve created that says, “We’re gonna assess the shadow IT. We’re gonna meet with everybody using shadow IT. We’re going to understand what they’re trying to do with it.
We’re going to identify the risks, and then we’re gonna put a s- a proposal together to help them utilize AI safely and [00:22:00] reduce the risk by enforcing policy.” Enforcing policy, but we have to know what they’re using, and we have to be able to measure it Where we do need a tool, but we’re not pitching the tool.
So review the top client accounts, have meetings with the first set, the next set, et cetera. Discuss those next priorities now. If it’s not an AI conversation, you can skew a little bit outside of it if you haven’t had a QBR with a client in a while, Rich. This might be a way for you to expand the conversation a little bit.
But ideally, it’s more of an AI conversation because we sense, statistics tell us, that the majority of SMBs out there are leveraging AI, and if the leadership doesn’t know it, it’s because the users aren’t telling them, right? ‘Cause they’re trying to stay outside, they’re trying to stay out of trouble.
And then connect every of one of those recommendations For each user or business unit that’s using AI to a business outcome, right? So we’re not [00:23:00] just coming in and selling, a, we’re gonna a project, but we’re selling a business outcome. So don’t tie it into, oh, once we’re done, like I used to in the early days.
Once we’re done, all your users will, will be able to send and receive mail safely because we put antivirus in. That’s not it. All of your users should be more efficient and productive because we’re helping them utilize AI properly. We’re building potentially agents and workflows and helping them maybe orchestrate, like crawl, walk, run, so that the…
So over time, as you’re building this 12-month visionary roadmap, for everything that they’re doing, you have a roadmap for AI. So h- here’s what the first 90 days looks like, here’s what the next 90 days look like, all the way through the year, so the client knows where you’re going next, and you have that buy-in, and it creates a tremendous amount of stickiness and loyalty and referral opportunities because you’re actually helping that, that business achieve its [00:24:00] business outcomes, not simply paying bills for technology support.
Rich: Surprises can be fun, Erick, but in the business context, they quite often aren’t, and the worst kind of surprises are AI and security surprises. Ask the 98% of organizations that ha-ha- have had an agent-related disruption, not a pleasant surprise. Yeah, absolutely getting out in front of both AI and security requirements in terms of the good that AI, the power AI can bring to a business, but also to head off the dangers from a security perspective it c- it can pose.
This is gonna be the kind of stuff that creates opportunity for you, builds loyalty for you, creates opportunity for the customer as well. The other surprise you want to avoid, though, as an MSP, is the surprise that comes when you discover that your customer is talking to somebody else about AI and security.
And I know just from conversations with MSPs that I’ve written about in my blog, Channel Hawk, [00:25:00] just within the last few weeks that the- there’s a lot of incoming traffic to MSPs right now from SMBs who really wanna have a serious conversation about AI and agents and getting real productivity-boosting impact out of those.
And so if there is that kind of interest and demand out there, if it is expanding as anecdotally and quantitatively, I know it is right now. That means you really need to be having that conversation and building that roadmap with your customers right now before somebody else gets in there to do that, which would be a really ugly surprise.
Erick: Absolutely right, Rich. And, that’s letting the wolf in the henhouse as as the adage goes. And, the other thing about creating this roadmap also is giving the client, getting their buy-in for budgeting over time, right? We’re not asking for, an emergency, fund to fix something that’s broken.
We’re saying, “Hey, these are risks. Let’s be strategic about this, and let’s break it out.” That’s why I mentioned that [00:26:00] 90-day kind of sprint over time. It’s like the first 90 days, here’s what we’re gonna do. Here’s the budget that we’re gonna need. Here’s the next 90 days, probably in this range, but it’ll depend on what we come, what findings result from this first 90 days, right?
And each one of those conversations then helps de-risk the next budget request and justify the next budget request, right? So the client sees where we’re going. This is the objective that I’ve shared with my MSP. This is what they’re helping me achieve, and this is how we’re gonna eat that elephant in bite-sized chunks so I can start planning ahead to budget.
And when you have a plan like that’s already been conceptually agreed to, to your point, Rich, along with phasing and budget, it’s gonna be really hard for somebody else to come knock on the door and kind of in- insert themselves into that conversation and try to displace you.
Rich: Okay. N-ABLE is getting deeper into AI.
We know that just within the last couple of months, Kaseya and ConnectWise have both had major AI related [00:27:00] announcements. But y- you’ll have trouble finding anyone who is more immersed in the productivity potential of AI for MSPs than somebody running an AI native startup that does exactly that.
And so our next guest actually is exactly such a person. His name is James Ferro. He’s the CEO and co-founder of a company called Cyft, C-Y-F-T. This is a service desk automation, but beyond kind of tool for MSPs. And we’re gonna get his perspective on the opportunity for MSPs out there with AI coming your way right after the break, folks, so stick around.
We will be right back
And welcome back to part two of this episode of the MSP Chat Podcast, where we are extremely pleased to be joined by the founder and CEO of Cyft. That is C-Y-F-T, an organization I have written about multiple times on my blog, Channel Hawk. And it is great to have [00:28:00] him on the show. His name is James Farrow.
James, welcome.
James: Yeah. Thank you. Thank you, Rich, and nice to see you, Erick.
Erick: Likewise.
James: Appreciate it.
Rich: Like I said I’ve covered Cyft. We’ve spoken in the past, but there are gonna be some people in the audience here who are new to you and new to the company. So just to kick things off, tell them a little bit about who you are and what Cyft is all about.
James: Absolutely. First and foremost, I am just a dude trying to figure it out like everybody else. But I’m able to spend a lot more time doing it, right? I’m not running an entire MSP. This is everything that I do all day long. It’s like I probably have bags under my eyes. It’s 14, 15 hours a day running the business by day and then learning and growing and doing engineering at night.
So I’ve spent a lot of time just with artificial intelligence, with LLMs, just spending time building with them manipulating them, getting them to do unique things. And I think that’s what I can bring to the table today, is everything that I’ve learned through trial and error, [00:29:00] through hundreds and hundreds of hours as well as what our customers are doing with with our product, Cyft
Rich: And just so folks understand, so what l- what your customers are doing with Sift right now, and w- we’ll talk a little bit in the context of where you see AI tooling for MSPs going.
Right now the Sift platform is helping them with documentation. It’s basically listening in on every technician interaction with the end user, and then instead of the technician having to capture all of that detail and documentation, Sift is doing all that for them and then building up a store of data along the way that will have some useful applications later on.
But that’s a pretty good summary of where you are right here and now, James.
James: That’s pretty great. I in a nutshell, we solve garbage in, garbage out. We sit on the technician’s machine, we hear everything. It doesn’t get translated through a human. We let machines talk to machines, which means you get very high fidelity signal into your system of record, [00:30:00] and then we query that for you using natural language.
So you can ask it really anything that you want. For the first time in history, you can actually get what you want out of Connectwise Manage just using n- natural language. So it’s really garbage in, garbage out, and we do have big plans for what we wanna do and things like that. But at its core that’s a huge problem, especially when we’re gearing up for an automated future where, we’re in the, I’d say we’re in the first third of it, and you need clean data.
You need to be able to query that data, and Sift just makes that so easy, and makes it really easy for the technicians to do all their documentation, and it’s so much better than if they did it themselves. So the business is pumped because they’re getting time entries on time, and they’re getting internal notes and resolution flags.
Everything looks really pretty, and every tech, every ticket, every time is very consistent. And it’s a boring problem to solve, but I think it’s the root problem that is super important for MSPs to be taking a look at
Rich: The, a- and the solution we’ve been talking about here has been GA for just a few months at this point.
So you’re in the process now of [00:31:00] talking regularly with MSPs about what you do and about what AI can do for an MSP, which puts you in a great position to kinda give us a lay of the land on AI-based automation among MSPs right now. To, to what extent and I’m asking you to generalize a little bit here.
Sure. To what extent, in your experience, are MSPs kinda hungry for the s- sort of thing that a company like Sift does, and looking for help and guidance? Or to what extent are you kinda evangelizing what AI can do for an MSP? And are there any patterns that sorta separate the early adopters from the companies that are maybe skeptical or laggards?
James: That’s a great question. I would break it up into a few different groups. You got the folks who are like, “What are you doing with my data?” Very security-conscious, right? They’re thinking about compliance, and they’re approaching it with that level of cautiousness. I applaud those people. I think that’s super important.
Then you got the [00:32:00] DIY folks, the, “I’ve got MCPs. They’re all plugged into Cloud Code, and they’re accessing my actual environment.” Now, if you’re able to lock it down, you’ve spent the time to do it correctly, that’s awesome. I met with the a really talented MSP team, and they built basically Sift internally for themselves.
Now, it cost them two engineers and a half a million dollars, and they still have to maintain it forward. But they were able to accomplish some of, if not all of what we did, but more tailored for themselves which was really fascinating. They were an outlier. Usually it’s I have a MCP to connect-wise, and I can do read/write through Cloud Code, and that’s that bucket of DIY.
And then I see the folks who are hungry for it and confused by what Sift does or Thread does or Pia or Zofik, and the market isn’t really sure, do I need [00:33:00] all of these? Do I need one of them? How do they talk to each other? Do they integrate? Do they work together? Nikhil at Neo Agents, shout him out too.
So there’s that bucket. And then you’ve got the just perfect fit for Sift, which is I need AI, I’m not gonna build it myself, and- I understand that the data really is the first step that I should be taking, is getting really good clean data. Before I start automating all this stuff, before I start putting agents in, let me get high fidelity signal from actual interactions into my PSA.
That’s the bucket that we play really well in, and that breaks down into large organizations with 200 plus technicians all the way down to the three-person team, and we cover all of those bases. The product is a really great fit for all of them
Erick: So James, we know that there’s a lot of, I will say promotion/noise in the channel- from everybody and their grandma about, the AI and the value of M- [00:34:00] AI to MSPs, to their customers and things like that. It c- it’s creating a lot of confusion, right? So what would you say to MSPs that are interested but haven’t really figured out how to take advantage of it?
They haven’t yet adopted a solution. You mentioned the first step is getting that signal clarity and purity. What kind of… how do, how would they do that? What kind of barriers should they expect to encounter, and what would you guide them to do to overcome those barriers, in this in a first step a process?
James: If I were an MSP, I would find a small subset of my team and dedicate them to trialing multiple different products. I wouldn’t be thinking about mass rollouts. I wouldn’t be thinking about multi-year agreements. I would be thinking about a dedicated team, people that I know can handle the added [00:35:00] bandwidth of learning a new tool, maybe, a- as long as they’re excited by it, and I would be trialing products right now in isolation, probably not even on my production environment, on testing staging instances, and having people try to break them, test them, and give feedback to founders like me.
That is exactly what I would be doing. I wouldn’t feel behind because I haven’t had this massive AI rollout. I think it’s really early. I think the founders like myself, the founders like Nikhil, like Pinar at Lexal, we are still figuring it out. The founders like Sam Altman, right? Those guys are still figuring it out.
So I, I would encourage people to not feel like you’re behind. The only way you’re behind is if you’re not exploring, and you don’t have to buy products, you don’t have to implement them, and roll them out, and do all that sort of stuff in order to explore. But I do encourage them to explore and to give feedback to founders like me so that we can have a really awesome future as we continue building.
That’s what I w- would recommend.
Erick: The thought that [00:36:00] came into my mind as I’m listening to you describe, your guidance to MSPs is, so you’re asking the MSPs to kind of Lewis and Clark it a little bit, right? Yeah. Lewis and Clarking your way through AI because you gotta- Yeah … assign a team, you give them the destination, and you’re like, “Figure it out along the way.”
James: And don’t feel like you’re behind right now. We are in an echo chamber on LinkedIn and all these sorts of things. You’re not behind if you aren’t implementing AI today. You will feel behind if you go on LinkedIn and all these posts and stuff like that. I’m gonna be one of them that’s posting and saying AI is the best thing in the world, of course.
But you should be strongly considering trials with companies, looking at it as development partnerships, things like that, understanding these aren’t gonna work perfectly. But that would be my recommendation. If we can get more MSPs on that kinda train of thinking, where it’s not, “I need to roll this whole thing out, and I’m, in such a rush to do it, and I need to find the next thing, and which company, how do these all work together?”
And it’s we could just slow down and just work [00:37:00] together as founders in the community and build something people actually want, that’s safe to use, that technicians love, that actually helps the business. That’s what we want, and we don’t need to be in a rush to do that.
Rich: It’s probably worth pointing out when you were talking about Pinar, that’s Pinar Ormaci from Lexal.
She’s actually been on the show before too. Now I’m thinking fair is fair. I’ve gotta get Nikhil on the show to- Yeah …
James: round out the trio.
Erick: We
James: need
Erick: a panel now. We need a panel for one of
James: the shows.
Rich: Yeah. Yeah. Get everybody in there. Yeah. Maybe we need to do that down the road.
Erick: Yeah.
Rich: One of the things that I really wanted to get into with you, and I was kinda hinting a little bit at the fact that y- you’re automating, like you said, this big pain point, around data and data cleanliness and documentation.
What you’re doing right now is actually very valuable for MSPs. But, there, there are documentation solutions. There’s service desk automation. Part of what I’m interested in is what comes after that? If that, we’re in early innings, that’s where [00:38:00] MSPs can get started. As you look down the road a little bit, ’cause this is something you’re thinking about to a much greater degree than most MSPs are, than, and than Erick and I are.
Where do you see AI powered automation for MSPs kinda going over the course of the next, let’s say, two, three years?
James: I’m gonna give you my best guess, and two or three years from now, if this clip surfaces, I’m probably gonna be way wrong. But I think automation and agentic orchestration and really agent runtime in general is gonna get commoditized. I think it’s gonna get easier and easier for MSPs to deploy it inside their own environments Which is going to set up a future where they’re now really thinking about how they can deploy in their customer environments.
Now, the reason why I think there’s going to be a huge shift in the [00:39:00] MSP business model isn’t the AI itself, it’s risk and liability. So when an end client, a 50-person law firm wants to implement AI, that’s a ton of risk to take on for themselves. If they have an AI that goes rogue and does something that hurts a client and they get sued, that might collapse their entire business.
So just like cybersecurity, they’re gonna need to pass that risk to a third party. I believe the MSPs are perfectly positioned to take on and absorb that risk. It’s going to change the underlying unit economics of an MSP. We’re talking about 70, 80% margins on that, where the MSP is actually implementing the AI in the end client’s environment with the caveat of, “I will absorb the risk.”
Now, if the MSPs are gonna be able to do that, they need verified work attestation essentially. So did my technician do the right thing? Can I [00:40:00] trace back this agent action all the way to a human decision that was made on my service desk based on our runbook? If you have that pathway, that audit trail all the way back, then the insurance carriers can underwrite that risk for the MSP.
So now we have this really awesome three-layer sort of AI liability and risk chain that is finally connected. The insurance carriers can’t underwrite something that they can’t prove and can’t see evidence for. The MSPs absolutely can’t do that themselves right now. They’re gonna need someone that’s able to do that.
And then if the end clients wanna deliver AI, they’re gonna need to pass that risk somewhere. So I think that’s really what the next maybe four to seven years is gonna move towards. We’re gonna see more outcome-based billing for the first time. The only reason that will exist is because you can verify the work.
So I really think it’s gonna be about mirroring the human and machine [00:41:00] decision-making in a way that we can prove. And if we can prove it, we can underwrite it. If we can underwrite it, everybody’s happy because there’s multiple throats to choke, and so you can do what you need to do as a, as an end client.
Rich: I, I’ll just jump in real quickly here Erick, if that’s okay, ’cause that was a re- so first of all very interesting answer. And it calls two things to mind for me. I wrote a post in Channel Hawk a few months ago about, the SAS apocalypse fears and w- what’s gonna… And I was writing more about vendors, but what’s gonna save vendors from- an end user or an MSP just doing everything through the L- LLM, what are they gonna…
And I identified a few factors that, are gonna be an answer that one of which was risk. And y- it’s exactly what you were talking about there, where maybe you can do it just you and Claude, but that’s gonna involve absorbing a lot of risk, and people are not gonna wanna do- So getting close to risk, the, I forget who it was who came up with this phrase, but embracing [00:42:00] risk is actually a good thing for vendors and for for for MSPs as well. And then the other thing I’ll just quickly say is just an hour ago, I was interviewing the founder of a company called Gentic Flow, and he was talking to me about the proof layer that he’s building into this.
And you’ve just clarified for me to a greater extent than I gave him time to why he’s doing that and why it’s so important. And it’s okay, got it. Got it. Now I understand. Interesting James.
James: Yeah. Yeah, I think it’s gonna be really important. I hope that Sift can be a part of that solution, for sure.
Erick: Yeah. I also wanna follow up on your answer because so much came out of that response. And this is what’s great about Rich and I, we sit on kind of a couple of different sides of the, channel from our thinking perspective. And the things that I wanna highlight were, number one, it sounds like you solved the how do I monetize AI problem for MSPs.
It’s like they’re trying to figure out how to package a pr- and do it. They’re not ready yet for their [00:43:00] customers ’cause they still gotta figure out for themselves. But what you just described, it’s, it feels like a very natural pathway now- … MSPs to now monetize and take that r- de-risk- aI for their clients. I like that a lot, James. The second thing, That struck me was, this idea that insurance carriers now can insure that risk. And what struck me was that pathway is already laid now for MSPs and vendors. There are vendors out there that say, “If you sell our stack, we’ve already partnered with such and such insurance company.
That stack now is insured, and you can go and, warranty your services basically for that customer.” So there is already a pathway now. It seems to me that it would make it easier for them to see AI and this opportunity to say, “Oh, if you’re doing it this way and you’re using, these best practices and these platforms, then yes, we will give you more favorable rates because it is proven and you can demonstrate that-
Through the [00:44:00] output.” So- Exactly … I wanted to highlight those two things.
James: Yep. That’s exactly right. Outcome-based billing, and then AI liability and risk absorption. Those will be- Oh, yeah … two things that’ll change, I think.
Erick: No, appreciate that. Sorry, didn’t mean to step on you. So I’m gonna ask you a Star Trek related question now because Star Trek, at least the version, when I was growing up, it was the original series, right?
And the thing that was most amazing to everybody watching that show was people were talking to the computer, and say, “Computer, do this or that,” and the computer would respond, blah, blah, blah. Do you see MSPs like the next evolution of AI, of them to get away from having to, chatbot this thing to death and just be able to have a natural language conversation?
I think we’re seeing early kinda signs of that, but it’s pretty rudimentary now. But do you see the ability for AI now to not only, support documenting and demonstrating, activity and workflows and things like that in [00:45:00] helping solve some of these level and level two kind of services?
But taking more action and then just being more of a conversation like we’re having with, I don’t wanna say the names of these AIs that are in everybody’s houses ’cause mine will start talking to me right now, if yeah. But is that kind of what you’re s- what you anticipate as it being more of a natural conversation that we can have and, by the way, I’ve got my phone and I’m in the car, and I can start working on somebody’s issues just by talking to my AI and having it, solve issues for us?
What are your thoughts?
James: Big question
Erick: Or is it way beyond that?
James: There’s kind of two parts in there. There’s, like, how are we gonna communicate with machines? And then there’s the where are we at with the action layer, I’ll say, and then maybe how do those two things work together.
So I’ll try to answer that in a cohesive way. The first part on communication, I don’t think chat goes away. That would be like Microsoft Teams chat going away or Slack. We [00:46:00] love chat. We chat all the time. It’s just a faster way to write a letter like we, did hundreds of years ago, right?
Written communication is always gonna be a first-class experience. People want chat to go away. They don’t. If you just take chat away, it sucks. Chat’s actually great. Maybe that’s controversial in 2026, but I think chat’s here to stay. Why would it not be? It’s just such a quick way to send a message to someone.
It’s just written communication.
I think it’s gonna be an and, right? You have chat and you can dictate right into it. Now, I don’t even think that has to be conversational, right? I think just the ability to chat and or dictate and have it go to text and put that in, it’s just faster than typing.
I think people care about speed. I don’t know that we’ll get to a two-way conversational back and forth. We can do it. I do it every day. I have built a bunch of voice bots and things like that, but it’s actually so much easier to just read, and if we think about human communication our [00:47:00] comprehension is much higher when we read, and our ability to communicate is much better when we talk.
And so voice in, text out, I think is gonna be the first-class workflow. Now you’ll still have and, right? So it’s just gonna be multimodal. So you can chat, you can dictate, you can do conversation mode. All those things will be available but I don’t think any single one of them just goes away over time.
So that was the first part on just communication. Then the s- second thing is on the action layer. I think we’re racing to try to do something that we don’t fully understand yet. Which is how do I automate all of these things that I don’t even know how my people are already doing for the most part?
And if I already know how the, my people are doing them and they’re documented, why would I not just use RPA? What is the value of agentic orchestration today for an MSP? Like today’s version. I don’t think there really is, because we don’t understand the workflows well [00:48:00] enough yet for the things that we would actually utilize it for.
Now, I can give you examples of things that we’re discussing at SIFF that might be helpful, like batch ticket processing after hours, right? Using LLMs to go back and look at your last six months of tickets and, fix those. Hey, maybe you’re moving to true methods or you want everything to be C-level.
You can go retroactively clean things up and use reasoning to look at the problem and recategorize it so you have cleaner data historically, things like that. Those are like very tight use cases for agentic orchestration. I don’t see this whole action layer of like agents doing a ton of this stuff today because if I’m an MSP, I’m probably just gonna use RPA for that thing ’cause I want it to happen the same way every time until I have an agent that’s just so good that I don’t need it to.
But I think there’s a big gap between can it do something really simple most of the time, [00:49:00] and I’m actually using agents to solve problems that are novel. So I think we’re farther away from that than maybe we think. Even though the technology’s there, I don’t think the workflows and the understanding and the real world chaotic service desk is ready for an agent to just start doing things.
And when you get Nikhil on here, he’ll have a totally different opinion that’s probably even more informed than mine. But that’s just how I feel about it today. I think we’re a little bit w- a little ways off.
Rich: Interesting. Interesting. So y- what you guys have built and are continuing to build is a system of action.
You kinda referred to how you connect with an MSP system of record right? And now I I recently wrote a piece about how N-able has rolled out an MCP server so that systems- … of action like Sifts can connect to that back end. And I saw you on LinkedIn thanking them for the vision- to make that possible, Yeah … for vendors who wanna do it and MSPs who wanna take advantage of [00:50:00] that. A- as you look down the road two, three years, and I’m not trying to set you up with a loaded question here, I’m just really kinda curious, a- as you think about the N-ables and the Kaseyas and the Connectwises of the world, is there a long-term role in the tool stack and the ecosystem to be the sort of back-end system of record f- for systems of action like Sifts to interact with, and they’ll be a little disintermediated from the MSP?
Or do you see th- ’cause they’re obviously building their own system of action functionality or acquiring it. Do you see them playing a combination of roles? What are the… How do you see the future playing out for those guys?
James: They’ve got a lot of money, so it’s tough to count them out in any sort of, significant way.
I think their products as they exist today are gonna have to change dramatically. That’s probably gonna happen through acquisition or through, like micro-acquisition acqui-hires. Bringing on a team like Mark and myself into [00:51:00] an organization like Kaseya and saying, “You guys just do, like research and development, and the cool things that you guys build and test with beta customers we’ll port into our product.”
If I was the CEO of Kaseya that would be how I would be thinking about it. Who’s doing cool stuff, and let me, like acqui-hire that and leave them alone and just keep doing cool stuff because it’s moving so fast. I just want people working on that internally. Maybe you hire that and, I don’t know.
But I would be thinking really about talent acquisition, and more than, like the product landscape today if I was those companies. They’re… You almost gotta just start fresh. You have your system of record, but like you gotta be thinking about what’s our next product. If I’m ConnectWise, Sift is ConnectWise only right now, right?
When we have the internal integrations for Halo and Autotask, so we can roll those out. But if I’m ConnectWise, I’m looking at a Sift and going how can I help Sift win because they’re making my product more valuable?” Like, how can I be a part of [00:52:00] that? That would be how I’m thinking about it, and I just don’t think vendors who are, that large really ha- even have that level of bandwidth.
They’re just very… there’s this huge mountain of, red tape and junk that they have to be working on and thinking about. They’re not really thinking about, which founders are maybe doing cool stuff. I don’t know if they I don’t know what they look like, the shape of them.
I think it’s gonna be acquisition city. I think it’s gonna be acquisition city, and if they really wanna rock and roll, they would leave the brands as they are for as long as possible, and the main reason is pricing power. Someone like Sift is gonna be able to charge a heck of a lot more for the exact same product than if you put ConnectWise on it or you put Kaseya on it, because we don’t have all of this historical precedent of, “It costs $40 or $50, and I would never pay more than that for this thing.”
So if I’m looking at it, I’m like, “Who can I acquire? Who can I… Who has a good brand that I can just put money [00:53:00] behind and put effort and fuel behind, keep the brand, let… introduce them to my biggest customers?” I’d almost do it an acquisition not even publicly, and I would just bolster that s- that startup so that they can reset the pricing power for AI for all my customers before I really announced it.
So I know that’s a little bit pie in the sky, but I don’t think that they can… I don’t know. I think they’re gonna have to do acquisition.
Rich: And I’ll just quickly say I… and I won’t call out the specific name, ’cause I can’t remember if I was on or off the record, but there is at least one really big company out there that is very quietly making acqui- And not vendor acquisitions necessarily, but they’re doing some of this acqui-hiring that you’re talking about James, where they encounter some people with really interesting skills and bring them on board.
James: That’s who I think’s gonna win. That is the formula, is you get weird, awesome, like hardworking people that are just obsessed and [00:54:00] put them in a sandbox, and anything cool that comes out of that utilize that. Looking for novel approaches to these things. It’s blue ocean. There’s, there actually is unique things that are coming out that create competitive advantages.
Look at Mythos, right? Mythos came out. That wasn’t because we plugged in 16 more data centers and just charged more power in- into it. It was a different technique that led to an emergent outcome. That’s a small team, probably led by one person, that is able to completely transform the landscape of how we think about it.
That still exists right now all over the place, on the application layer, on the model layer, on the data layer, on the memory layer, on how we think about skills, context loading, caching. There’s so much for really brilliant people to sink their teeth into and create competitive advantages for products.
It’s not just the foundational models. It’s the harness around it that is [00:55:00] really the opportunity, in my opinion
Erick: I was having a beer the other night with a good friend of mine, James, who’s so AI forward, his organization they’re… Like, it’s amazing. And he was telling me about this next level of communication, and he…
and, I didn’t understand it, but it came off to me like this is like command line interface between these LLMs and things like that. What… And he said it would reduce cost, you know- … token burn and things like, ’cause a more direct way. Let the machines talk to the machines the way they talk.
Do you have any thoughts about that?
James: I was just thinking about it. I was just thinking about it. It’s like MCPs are probably going away, but it’s tough because the market is just now starting to understand MCPs. So founders like myself and my co-founder, Mark we’re talking about these things every day.
We’re like, “Should we even build the MCP harness when we know CLI is actually the right move?” Skate where the [00:56:00] puck is going. But then it’s of course we gotta explain MCPs and why they’re inefficient. They send all these tools through instead of just letting the machines talk to each other.
Was really cool before, but now things are moving so fast. CLI is the future. It’s really actually difficult to decide what to do and how fast to move. But yeah, totally. Internally we know MCPs are, like, gonna be out the door from a technology standpoint, but they’re just now hitting adoption, so it’s weird.
Erick: That’s how fast things are moving. You’re predicting things maybe four to seven years out. I’m… I was thinking, “Man, that’s a long time to wait in in AI time.”
James: Yeah, absolutely.
Erick: What are your thoughts on, the… We know that there are a lot of there’s a lot of investment coming in to the market building out AI led MSPs, call them, AI frontier firms and things like that.
What are your thoughts around, that type of focus [00:57:00] and the gap between… potential gap between those folks that are really leaning hard into AI to build out these MSP practices and investing in them and things like that, and the folks that are just not yet thinking about that?
What are you… what are your thoughts around the valuation, the, the interest that these AI led organizations are creating in the channel, and what does that mean for the rest of the market?
James: I think we’re gonna see mass consolidation. I think it’s gonna accelerate which I don’t think is necess- we kinda talk about a bad thing, but I’m like, a lot of the guys I talk to are like, “I’m five to seven years out. I’m just like, I’m gonna be selling.” And so I’m like, I don’t know. I see on LinkedIn and everyone’s it’s this, the whole thing, stop the consolid- I think this is half of what people want, is like accelerate that timeline.
Now, [00:58:00] what’s your multiple when you get there? That’s the thing that I have conversations around, is okay, you’re four years out and, the macro market is accelerating that timeline probably in, in a lot of ways. What’s gonna be your multiple? Are you gonna have really crappy data that you can’t prove, and you have no workflows documented and all this sort of thing?
And, kinda get to the due diligence table and they’re gonna ax your price a little bit once they dig into things. Or are you gonna have really clean data awesome workflows documented not a lot of key man risk around your customer service, like all those sorts of things, and have a much higher multiple?
And I think that’s probably what the conversation should be around when it comes to the consolidation piece. Now, what does the AI first MSP do for the market? It’s hard to say because you can slap AI on anything. I don’t know if the guys there and gals there are that smart. If they are brilliant, then holy crap, they’re [00:59:00] gonna have a massive opportunity.
If they’re not then they’re just gonna spend a bunch of money building things that they could have just bought for 1/100 of the price, and that every other MSP is just gonna end up buying anyways. So it, it really right now is the people. If you have smart people that are, like, pushing the limits and actually understand how to productize things and add value, that is the competitive advantage.
So when I see someone raise a bunch of money and, they, they have AI on it, it’s like, it’s hard to know. It’s really hard to know because it’s not that difficult to raise money, and until people are putting out products we don’t know. It’s so blue ocean, we’re so early on that piece that yeah, it’s hard to say.
I think it’ll separate over the next few years, and we’ll start to see a divergence of people that plan to build all this AI internally and didn’t have the chops, didn’t hire the right people, didn’t spend the egregious money on salaries for those people to get that competitive [01:00:00] advantage. Those people don’t work for other people for the most part.
The people that you actually want, they probably don’t work for someone else. They’re probably working on their own. That’s why I was talking about acqui-hire. It’s it– these guys aren’t just W-2 employees, right? For the most part, they’re either doing it themselves or they have a company or, they’re contractors, whatever.
But that was a long-winded answer to your question. But I got a lot of… I think about this a lot and what does that mean? It’s like it actually doesn’t mean anything if we don’t know who the people are. That’s the short version.
Rich: The fact that you’re thinking about this stuff a lot is why we wanted you on the show.
This has been really great stuff. W- last question for you. So if you are an MSP in our audience here who does want to be AI forward and kinda lean into what AI-powered automation can do from them, these are great times and very difficult times because there are a lot of options out there.
Like you were saying a smart thing people can do right now is just create a lab and start testing these [01:01:00] things. Give folks in the audience a little a- advice, any thoughts you can share in terms of how to run that eval process, meaning like what am I looking for? How am I comparing these tools?
H- how do I even approach making a decision on where I want to invest?
James: It’s a really great question
Where do you start? I’ve spent months telling people that Cyft is the place that you start. Start with the data. Clean your data up, something that’s easy to use, improve your time entries, improve the workflows that you kinda have right now and get better clean data in there. I do think that’s a great place to start.
So I’ll throw out Lexful too. Lexful’s doing great with documentation. So I’m gonna be looking at companies like Cyft, like Lexful, that are, like, not super intrusive into the things that you’re already doing now, that are just helping the things you’re doing now. Documentation is an awesome place to start, and if there’s another company that pops up that does exactly what Cyft does, I’m gonna promote them [01:02:00] too, because it is not just about Cyft or about Lexful or anyone, it’s just the whole ecosystem needs to be working together to help MSPs start to adopt this technology.
I would start with documentation and getting good clean signal into your system of record, because everything else is built on top of that. Every workflow that you run, every analysis that you do any sort of way that you’re looking at your data, if it’s broken on the input every downstream product that you buy is broken.
So fix the input, fix the capture layer. Right now I believe that’s Cyft. Lexful does a good job with that too, and so that’s where I would start.
Erick: All right.
Rich: James, we are appreciative that you were able to carve out some time and join us here on the show. Very interesting conversation.
Lots more that we could theoretically get into with you. For now, for the folks in our audience here who wanna get in touch with you, learn more about Cyft, where should they go?
James: Yeah. Cyft.ai, great place to start. That’s [01:03:00] C-Y-F-T dot A-I. And then you can find me on LinkedIn, James Farrow, F-A-R-R-O-W. I’m always happy to chat, take calls like this.
I think about this all day long, so I am here. Use me as a resource. For most people, I have really nothing to sell you unless you’re, like, on Connectwise and you’re looking for AI and you need document… unless you’re, like very specifically in our niche I have nothing to sell you, and I will tell you everything that I know, everything that I’m learning, and I am just here to help and be a conduit of the experience and wisdom that we’re gaining by building.
Rich: All right. I love it. Thank you, James, for joining us on the show. Folks, Erick and I are going to take a quick break now. When we come back on the other side, we’re gonna share some some final thoughts about this interesting conversation with James, have a little fun, wrap up the show. Stick around.
We’re gonna be right back.
Welcome back, part three of this episode of the MSP Chat Podcast. One last [01:04:00] thank you to James Farrow for taking a little bit of time out to share his thoughts with us. I’ve had multiple conversations over the course of the last year-ish at this point to speak with him and I was very glad to have him on the show ’cause he always has sort of an AI pioneer’s perspective on, on what’s going on with that technology and managed services right now.
Very thoughtful outlook and so it was great to get his point of view here on the show with us. I think and it’s interesting that the a- as an AI insider, which he is, I think it was interesting and should be encouraging in the context of everything we were talking about earlier in the show, that as he put it, we’re, in the early innings of the AI story in managed services.
And so nobody is too late at this point, or even on the verge of being now … things are moving very fast, and we’ve kinda talked about how you don’t have forever to get moving and figure some of this stuff out. But that said we’re early in the story. Y- you are not hopelessly behind if you feel like you’re still just getting things [01:05:00] underway.
This is a perfect time actually to to be getting things underway. It was interesting, there was a point in the conversation there where he was kinda, … we were talking earlier about N-able’s point of view on getting the AI through the same pane of glass that you’re getting a lot of your other managed services from has certain benefits for an MSP, certain efficiencies.
And obviously, ConnectWise and Kaseya and companies like those will share that perspective. And James I won’t say he shared that perspective either, but he, y- he didn’t dismiss it at, at all. Yeah he didn’t give you the point of view basically that said, “Those guys will never understand AI as well as
He was kinda talking about how companies like his, one of the things they’re gonna have to consider is looking for an acquisition partner. It’s gonna be har- It’ll be a competitive challenge, I think going forward for companies like Sift. And James knows that right now.
I’m sure his peers at a lot of these startups know that as well. So that was one of many interesting [01:06:00] things that we got from James.
Erick: Yeah, and the way you position it, Rich, having the perspective of an AI pioneer is very different. The conversation is different, the viewpoints are different than, other folks that are out there just have been in the bus- in the industry for a long time and are trying to add AI into their existing solutions or platforms and things like that.
When you’re building it from the ground up with a very specific purpose and seeing I think you see the potential and the obstacles a little bit differently. I’m not saying that folks that are trying to integrate it into, their existing ecosystem don’t see the obstacles, but I think they’re, the pioneers are solving for different types of obstacles and thinking much further beyond, managing, an SMB’s AI usage, right?
The AI, the data is the key, right? And what can be done with that data. So it’s interesting to have, someone like James come in and [01:07:00] give us, his insights, and it really got me thinking, in, in a different way about a lot of things at, in front of us here in the next three to five years.
And, I also think that it’s not gonna take five years for a lot of this stuff to… I- it’s just happening so fast it’s hard to stay on top of all the changes and all the new entrants into the industry and the new solutions. And some of the folks that are being acquired you know- didn’t build their solution for the IT industry, right?
So that’s also interesting is we’re seeing these strategic decisions that the acquirers are making. So very interesting conversation.
Rich: And, you remind me, you, you talk about how quickly things are moving in this industry, and just one quick little example is just within the last few hours as we’re recording this access to Fable 5, the top-of-the-line model from Anthropic has been restored.
And our guest, our interview guest on next week’s [01:08:00] show is the chief information security officer from Barracuda, and our discussion topic is going to be these leading-edge models, Mythos and and Fable. That’s an interview that I booked within the last few days before I knew that it was gonna be exactly as relevant as it turns out to be.
And it’s just one, one good example a- of how quickly things are moving out there and why it’s so important to be talking about this stuff with your customers right now. But for now Erick, we have time for just one last thing, and it comes to us, of course from the world of of AI and specifically from Arizona, where a, the police were called to investigate a a potential shooting incident I believe in Chandler, Arizona and they called out the SWAT team.
SWAT team was doing that very dangerous work when they were, to their surprise, joined by a robotic assistant who was not technically part of the SWAT team. This techni- this robotic assistant is named [01:09:00] Dot. Dot works for DoorDash and apparently somebody I- maybe it was a member of the SWAT team, maybe it was somebody inside the home there, but somebody was in the mood for pizza and Dot came by to deliver.
Fortunately Dot was not hurt. Dot did not result in anyone else getting hurt here. Apparently, DoorDash had to send a van out to this house and I guess there’s video online that I haven’t seen of somebody from DoorDash with a controller in his hands kinda telling Dot, “Okay, c- come on back home here.
This is not a good place for you to be making a delivery.” But you gotta applaud Dot’s persistence. If it’s me and I’m a pizza delivery guy and I see a bunch of people in, in tactical gear with long weapons pointing them at a house, eh, I’m probably just gonna back off. But Dot was gonna get this done.
Erick: Oh, my goodness. I just had a flashback, Rich, of, like, all of these hostage movies we’ve seen where they say, “Okay, send pizza in,” and, the delivery guy’s in there, but he’s actually [01:10:00] the, the double agent that’s gonna, help free the hostages. I don’t think Dot’s got those kinds of skills.
Rich: Folks, that is all the time we’ve got for you this week on the show. We thank you very much for joining us. We’re gonna be back in a week’s time with another episode for you. I’ll just remind you until then that this is both a video and an audio podcast, which means that if you’re listening to us right now but you’d like to check us out on video, you can go to YouTube, look up MSP Chat there.
If you are watching us on YouTube but you’re into audio podcasts, go to Spotify, Google, Apple, wherever you get them, and you’re gonna find us there, too. And wherever you find us, please subscribe, rate, review. It’s gonna help other people find and enjoy the show just like you do. This show is produced by the great Riley Simpson who I attended a baseball game with earlier in the week.
He’s just part of the team with us here at Channel Mastered, where we work with MSPs looking to build… or excuse me, with vendors looking to build thriving MSP channels. You can learn all about the many ways we do that at [01:11:00] our website, which is www.channelmastered.com. Channel Mastered has a sister organization called MSP Mastered.
That’s Erick and his team working one-on-one with MSPs to help them grow and optimize their business. You can learn more about that at www.mspmastered.com. So once again, we thank you for joining us. We’ll see you in a week. Until then, please remember, as we always urge you to, you can’t spell channel without [01:12:00] M-S-P.
Check out more episodes
