Three experienced MSP operators tackle the most common claims about AI in managed services, separating real-world use cases from hype, and helping MSPs understand where AI genuinely helps, where it falls short, and how to evaluate and adopt it responsibly.
Key Topics Covered
- Whether MSPs need AI to deliver competitive services in 2024 and beyond
- The reality of AI replacing human technicians at MSPs
- AI accessibility for small and mid-sized MSPs versus large enterprises
- How AI goes beyond basic ticket routing and automation
- The risk of AI leading to increased errors and downtime without proper oversight
- Whether AI integration is as complex and disruptive as commonly assumed
- The true cost of AI adoption for MSPs today
- Data privacy and cybersecurity risks from AI tools in regulated environments
- How to measure ROI on AI investments
- How AI improves MSP security capabilities
- A framework for evaluating and selecting AI tools
Fact or Fiction: MSPs Don’t Need AI to Deliver Services Effectively
The panel agreed this is partly true and mostly beside the point. You do not need AI to run an MSP, but the more useful question is whether AI enables a meaningfully better level of service.
Michael Perrinich framed it as a knowledge-sharing problem. Individual technicians accumulate deep knowledge of their clients’ environments over time, but compiling and transferring that knowledge across a team is hard. A well-implemented AI tool that can read historical tickets, surface documentation, and suggest solutions based on similar past issues adds to the technician’s effectiveness in the moment, without replacing the relationship.
Jasper Grewal agreed, adding that the risk of AI is in using it inappropriately. MSP businesses are built on personal relationships, and handing first-line support to AI is a line he would not cross. The technicians at ROI Technology know their clients’ business models, workflows, and personalities, and no amount of training data can replicate that.
Andrew Moon took a broader view. The real value of AI for MSPs is leverage: tools that allow you to do more with the same headcount, grow a scalable business, and free up time for the relationship-building that actually retains clients. For Andrew, AI is a differentiator MSPs can take to their end customers as well, helping clients get more productive with the technology rather than just selling a support contract.
Fact or Fiction: AI Will Replace Human Technicians in MSPs
All three panelists rejected this firmly. No current or near-term AI model will replace the human technician, particularly at the relationship layer that defines MSP value.
Michael Perrinich pointed to the knowledge gap in working with non-technical end users. MSPs serve regular businesses, not tech companies. Their end users are not comfortable clicking into settings or trusting an AI to tell them a change is safe. The confidence and calm that a human technician provides when something breaks is not something any AI can convincingly replicate.
Jasper Grewal added that the contextual knowledge of individual clients, who is out on Wednesdays, which users are nervous about changes, whose business has just gone through a transition, is not capturable in a ticket system or a training dataset. That knowledge lives in long-term relationships.
Andrew Moon offered the contrarian position: he hopes AI can eliminate tier-one ticket volume, specifically the low-level, repetitive work like password resets and printer issues that consumes junior technician time and prevents senior staff from doing higher-value work. He was careful to note that this is not about replacing people but about changing what they are hired to do. If AI handles the minutiae, MSPs can hire at a higher skill level from the start.
Jasper pushed back with a fair point: many of these tier-one problems can already be solved with self-service portals and scripting that MSPs have had access to for years. The challenge Andrew identified is real, though: most MSP owners do not take the time to build those systems. Andrew’s hope is that AI can lower the barrier to that kind of proactive improvement.
The group landed on a middle ground: AI as a triage and suggestion layer that gives technicians better information faster, while humans own the client relationship and the final judgment.
Fact or Fiction: AI-Driven MSPs Are Only for Large Enterprises
False. The panelists were unanimous.
Michael Perrinich noted that tools like Syncro are actively building AI into their PSA platform at no additional cost, which means small MSPs get the benefit without an enterprise procurement process. The landscape of accessible AI tools has expanded rapidly: ChatGPT, Copilot, AWS Comprehend, Google Gemini, and others are available to anyone with a subscription.
Jasper Grewal pointed to Microsoft Copilot as a specific example of a free, useful tool for any MSP. For navigating Microsoft documentation, he called it the best tool available, full stop.
Andrew Moon noted that AI has followed the same adoption curve as every other technology: early adoption was expensive and enterprise-only, but the novelty and cost premium are gone. AI is now built into almost every tool MSPs already pay for.
Fact or Fiction: AI in MSPs Handles Only Basic Automation
False. Ticket routing and keyword-based field updates are the floor, not the ceiling.
Michael Perrinich described the meaningful difference between legacy automation and AI-powered automation: traditional systems matched keywords and changed fields. AI systems can read and interpret the full context of a ticket, understand what the user is actually describing, and surface relevant documentation or similar resolved issues. That is a qualitative difference, not just an incremental one.
Andrew Moon expanded the scope significantly. The number-one problem he hears from MSPs is sales and marketing, and AI removes the excuse for not doing it. Content creation, marketing plans, video scripts, and outreach sequences can all be generated at 98 percent completion using the right AI tools. For MSPs who struggle to find time to grow their business, that kind of leverage is transformative.
He also highlighted SOP development as a concrete, underused AI application. Sitting in a room with zero subject matter knowledge about a specific industry, Andrew used AI to generate a complete SOP for a sign company in real time, including load procedures, required tools, and OSHA guidelines. The audience, who were experts in the field, were impressed by the output. For MSPs who have been putting off SOP documentation for years, AI removes the activation energy for getting it done.
The conversation also turned to content quality. Jasper raised a fair concern: AI-generated content that is copy-pasted without editing creates a flood of similar, undifferentiated copy. Andrew acknowledged this and noted that the tools that work best for content are those that learn from your own voice and prior content, so that the output reflects your actual perspective rather than a generic template.
Fact or Fiction: AI in MSPs Leads to More Errors and Downtime
Conditionally true: if you do not apply oversight, yes. If you treat AI like any other team member and spot-check its work, no.
Michael Perrinich made the analogy plainly: you spot-check your technicians’ work, and you spot-check AI output. That is just the process of quality control. The risk is not AI itself but blind trust: someone who assumes AI will always be right and removes human review from the loop is setting up failures.
Jasper Grewal noted that the promise of AI, contextual awareness that can catch when something does not add up, is also its failure mode: it can produce confidently wrong output if the inputs are unclear or the model does not flag its own uncertainty.
Andrew Moon summarized it as “trust but verify,” a principle that applies to people, computers, and AI alike. For marketing content especially, the “garbage in, garbage out” risk is real. AI content that is not reviewed and personalized before publishing reflects poorly on the business.
Fact or Fiction: AI Integration Is Complex and Time-Consuming
It depends entirely on how the AI is delivered.
Michael Perrinich explained why he became more interested in AI after Syncro integrated it directly into the PSA: the vendor did the complexity work, and the feature arrived ready to use. By contrast, other platforms he has worked with required significant configuration before delivering any value. The right partnership reduces the complexity to near zero.
Jasper Grewal agreed: if you are using a tool like Syncro that has done the heavy lifting, you get the benefit from day one. If you are building AI workflows from the ground up, complexity and time cost are real.
Andrew Moon focused on SOPs as a practical example of AI removing complexity from one of the most time-consuming and consistently neglected MSP operations tasks. He described a vision for dynamic SOPs: as tickets are debriefed over time, the SOP updates automatically, and the most current version is available wherever the technician is doing the work. That kind of system, which would have required significant engineering investment in the past, is increasingly achievable with AI-assisted tools.
Fact or Fiction: AI-Driven MSP Services Are More Expensive
No longer true as a general statement.
Michael Perrinich noted that AI is now a standard feature in many tools MSPs already use, not an upsell. The compliance angle adds nuance: for MSPs serving HIPAA, FDA, or other regulated industries, you have to use AI tools from compliant providers, and those do carry a cost premium. At August eTech, that means using Microsoft’s HIPAA-compliant tools rather than general-purpose AI platforms, even for internal use.
Jasper Grewal extended this to regulated verticals generally: the cost of compliance-grade AI is no different than the cost of any other compliance-grade tool. You pass it along in your service pricing and maintain your margin. The cost is real but manageable.
Andrew Moon agreed that AI has commoditized. The risk of increased cost now comes from the human factor: the time spent training staff, governing usage, and reviewing output. That is not a tool cost, it is an operational cost, and it is manageable with the right processes.
Michael Perrinich raised one often-overlooked question: who is watching the watchmen? As AI handles more ticket research and documentation, someone needs to verify the outputs are accurate, the data is not creating bad recommendations, and the technicians are using the tools correctly. That oversight role is worth accounting for when calculating the true cost of AI adoption.
Data Privacy and Security Risks: The Real AI Concern for MSPs
This was the panel’s most clear-cut “true” answer.
Michael Perrinich shared a first-hand account of an employee who added an AI meeting recorder to company Teams calls without seeking authorization. The tool required no installation and simply joined scheduled meetings. When Michael reviewed the terms of service, he found that while the company claimed not to sell user data, their policy allowed any acquirer to do whatever they chose with data in the event of a sale or merger.
For a company serving HIPAA-regulated clients, this was a significant compliance risk. Michael escalated to C-level leadership, walked them through the terms of service, and the tool was immediately shut down. His takeaway: read the privacy policies of every AI tool before allowing it to touch your environment or your clients’ environments.
His guidance for MSPs in regulated industries: prefer tools from providers with clear compliance certifications (SOC 2, ISO, HIPAA-compliant data centers) and treat AI tool adoption with the same due diligence as any other vendor evaluation.
Andrew Moon framed this as one of the biggest opportunities in the MSP-client relationship. Clients trust their MSP as the expert, the surgeon who prevents problems rather than the janitor who cleans them up. A short educational video or client communication about which AI tools are safe for their industry, and which are not, is a practical, high-value touchpoint. LinkedIn’s default AI opt-in, which was turned on for all users without explicit notification, is another example of the kind of thing MSPs can flag proactively for their clients.
Jasper Grewal noted that shadow IT carries the same risk whether AI is involved or not. The bigger issue is unauthorized tool adoption broadly, and AI just raises the stakes because of the data these tools can access.
Product Features Covered in This Webinar
- Syncro Team Plan (referenced in sponsor offer: three months free with three-plus users on annual plan)
- Syncro PSA with native AI integration for ticketing and smart search
- Syncro HIPAA-compliant data environment for regulated MSP clients
- Microsoft Copilot as a built-in, no-cost AI tool for documentation search
- Microsoft Defender and Defender for Cloud with AI-powered threat detection
John K. Waters: Thanks, Victoria. Hi everyone. Welcome to today’s live tech talk, “MSP Masterclass: Navigating AI Facts and Fictions.” This event was organized by the folks at Redmond Channel Partner and sponsored by Syncro, provider of an all-in-one platform for managed service providers that allows them to monitor and manage client devices remotely, handle ticketing, and integrate billing, invoicing, and customer communication into one centralized system.
I’m John K. Waters, Editor-in-Chief for the Converge360 Group of 1105 Media, and I’m joined today by Jasper Grewal, business leader at ROI Technology; Andrew Moon, business and technology strategist at Orange Nomad; and Michael Perrinich, Director of Operations at August eTech.
A few housekeeping notes: this tech talk is being recorded. At the end, we’ll have a five-to-ten-minute Q&A. Please type your questions into the Q&A box as they occur to you. And Syncro has a special promotion for today’s attendees: three months free when you sign up three-plus users on the annual team plan.
Let’s have our panelists introduce themselves. Jasper, why don’t you start?
Jasper Grewal: Thanks. My name is Jasper. I’ve been in IT in some capacity for the majority of my adult life. My day-to-day is somewhere between DevOps and leadership, but I’ve done pretty much every IT job at some point, from desktop support through sysadmin, IT manager, to where I am now at ROI Technology. We’re a cybersecurity-focused MSP about an hour north of Seattle, primarily serving SMBs in and around Washington state. We’ve been around since 2014, and I’d say we’re pretty operationally mature. I heard a quote from a vendor recently that resonated with me: “Perfect is the enemy of good.” That’s pretty much our core business philosophy.
Andrew Moon: I’ve been an entrepreneur since eight years old. I’ve been in the IT channel since 1992, starting with Windows 3.1. I ran my own MSP until 2014 and semi-retired from day-to-day IT operations. These days I’ve fallen in love with live broadcasting, coaching, and training, and I spend most of my time helping small business owners run their businesses rather than having their businesses run them.
Michael Perrinich: My name is Michael Perrinich, from central New Jersey. I work for August eTech. I’ve been doing this for probably too long. My first computer was a TRS-80 in 1977. August eTech is a full MSP IT support company. We focus on biopharma and compliance, but we work with all types of compliance, including PCI, FDA, IRS, and anything else you can imagine. We also supply consulting CIO services. Our CEO is a consulting CIO for all of our client companies, which gives them a little extra edge. Our longest-running client is 21 years.
John K. Waters: Artificial intelligence is on everybody’s mind. Much has been said about AI’s ability to help MSPs automate processes, improve security, enhance customer support, and provide predictive maintenance. But the AI hype has been loud, and questions and misconceptions abound. So here’s how today works: I’m going to read some statements we’ve all heard about MSPs and AI, and you’re going to help us decide whether they are fact or fiction.
Claim 1: MSPs Don’t Need AI to Deliver Services Effectively
John K. Waters: “MSPs don’t need AI to deliver services effectively. They can deliver the same level of service without AI. It’s just a buzzword.” What do you think, Michael?
Michael Perrinich: We can absolutely deliver services without AI. But like any tool, AI can help us narrow down problems and focus on granular parts of an issue. Each technician on my team has their own individual knowledge. They know their clients’ tickets and the problems they’ve solved. That’s the easy part. Being able to compile that knowledge from all of your technicians and make it available to the next person is where the challenge is. A good AI tool that can read your tickets, consume information, and give you suggestions for potential problems is where it adds value. It can enhance the experience while you’re assisting your customers.
Is it necessary? No. Is it a buzzword? In some ways, sure. Honestly, I’m not sure “artificial intelligence” is even the right name for it. But as a tool for sharing knowledge and accessing documentation quickly, it’s fantastic.
Jasper Grewal: You don’t need it, but used appropriately, it can level up your team. But our businesses are built on relationships. The day that MSPs try to outsource first-line support to AI is going to be a bad day. My techs know their customers intimately: their business models, how they work, who does what. That knowledge doesn’t live in documentation.
Claim 2: AI Will Replace Human Technicians in MSPs
Andrew Moon: I think the key phrase here is “can they deliver the same level of service?” That’s really what MSPs have been searching for since the beginning. We’re looking for leverage. Tools that can give us more leverage so that we can grow a scalable, manageable business. AI is one of those game-changers, both internally and for your end customers. If you can teach your customers how to get their arms around AI and become more productive, that’s a differentiator you can put in your sales toolkit.
John K. Waters: “AI will replace human technicians in MSPs. The technology will eliminate the need for human IT professionals.” Michael?
Michael Perrinich: Not going to happen. AI can enhance the technician, not replace them. We all want our technicians to give customers their full attention. When a customer calls, we want to give them time, focus, and service above and beyond. Where AI comes in is as a smart tool. Yes, you can set up AI for simple things like password resets. But replace a human technician? I don’t think so. Especially when you’re dealing with external customers at various knowledge levels. We’re not supporting tech companies. We’re supporting regular people who don’t want to touch settings because they’re afraid of breaking something. AI can’t give a non-technical person the confidence to make a change. You’re never going to replace the human touch.
Jasper Grewal: Zero percent. Not in our lifetime. The most advanced models aren’t going to know that Joe is a fan of the local football team or that Patti works remotely on Wednesdays. The nuances of customer relationships are not replaceable with training data. I don’t want AI resetting passwords, because when something goes wrong, the warm bodies are the ones cleaning it up.
Andrew Moon: I’ll throw out a slightly contrarian view. I hope AI does eliminate level-one ticket volume. The number-one time drain I see when I coach MSPs is training and managing tier-one technicians who handle low-level issues like printers and email. If AI could handle some of that load, MSPs could divert that capacity to the high-value relationship work. Will it replace people? I doubt it. But I’d love to see it take some of the workload that prevents us from having the leverage to build real relationships.
Jasper Grewal: But do you need AI to replace a level one? Self-service password reset has been around for ages. Simple scripting handles a lot of those problems. You don’t really need AI for the low-level stuff.
Andrew Moon: You’re right. The challenge is that most MSP owners don’t take the time to build those systems. When I ran my MSP, I’d debrief with my techs every Friday, look at tickets, and ask: how do we eliminate these problems from coming in at all? That’s the work that never gets done because we’re too busy running. Can AI help filter through the noise and say, “Look at these four things”? That’s where I’d love to see it help.
Michael Perrinich: Right in the middle is where the answer is. Jasper, you’re right that some people will always want to talk to a person. But Andrew, you’re also right that tier-one is the biggest time drain. If AI could handle even a portion of it, what could your first-level techs do with that recaptured time? Training, right? There are two sides to it. But the relationships, knowing how someone’s kid’s graduation went, checking in after a vacation, those things will never be replaced.
Jasper Grewal: My fear is that if MSPs become too reliant on automated communication, it becomes impersonal. Reports go straight into the round file. The main touchpoint clients have with us is when they need help. We try to make that experience excellent every single time.
Claim 3: AI-Driven MSPs Are Only for Large Enterprises
John K. Waters: “AI-driven MSPs are only for large enterprises. They’re the only ones that can afford AI-driven MSP services because of the cost and complexity of integration.”
Michael Perrinich: False. There was probably a time when that was true, but not anymore. We’re already seeing AI tools rolling out in smaller services. Syncro is building AI into their platform right now. As more options come to market, the cost keeps dropping. And importantly, you don’t need the full ChatGPT experience to get AI-powered capabilities. There are specific components you can use. These tools are great for small companies, and they’re really not as expensive as people think.
Jasper Grewal: Absolutely not just for big companies. There’s no shortage of tools that even small MSPs can leverage. Claude for coding, Gemini, AWS Comprehend. All the major vendors have already rolled out versions that anyone can take advantage of. The devil’s in the details: it’s all about how you use it.
Andrew Moon: Millions of people are getting access to AI through every new Apple device. ChatGPT is twenty dollars a month. It’s built into every tool we already have. The early-adopter premium is gone.
Michael Perrinich: Copilot is in Windows 11 right now.
Jasper Grewal: Honestly, Copilot is the best way to search Microsoft documentation. If you’re looking for an answer in the Microsoft docs, there’s nothing better. It’s free, accessible, and genuinely useful.
Claim 4: AI in MSPs Handles Only Basic Automation
John K. Waters: “AI in MSPs is limited to simple automation tasks like ticket routing or basic system monitoring.”
Michael Perrinich: Couldn’t be more false. Basic automation has been around for twenty years. “Find a keyword, change a field.” AI actually levels that up: it can digest information, understand context, and give you more options. That’s where it really helps with ticket handling. Moving away from strict keywords toward contextual understanding is a genuine step forward.
Jasper Grewal: It’s going to depend on your operational maturity as an MSP. Ticket automation is the low-hanging fruit; that’s probably where most products will start. Whether you can use AI for more sophisticated things depends on how much you’ve developed your operations and documentation.
Andrew Moon: It goes way beyond operational basics. The number-one problem I hear from MSPs when I coach them is sales and marketing. AI removes the excuse for not doing it. You can create content, marketing plans, and outreach sequences at 98 percent completion with the right tools. That was the hardest problem for me when I ran my MSP: getting in front of more people. We have that capability now. And I look five levels deep: how can an attorney, or a property manager, use AI in their business to get closer to their goals? If MSPs can help their clients answer that question, that’s a completely different level of service conversation.
Jasper Grewal: Do you think the quality of content has suffered because too many people are copy-pasting AI output without editing it?
Andrew Moon: That’s where using a tool that learns your voice matters. When AI can draw on your existing content and the way you teach, the output reflects your actual perspective. The goal is 98 percent done, not 100 percent copied. The ideas are there; you still bring the judgment.
Michael Perrinich: The first step is the hardest. A lot of people see AI as a tool they don’t know how to use, and the fear of wasted time prevents them from investing in learning it. But once you understand what it can actually do for you, the initial time investment pays back fast. The key is picking one problem you want to solve and using the tool to solve it. That unlocks the fear of wasted time better than any overview or demo.
Claim 5: AI in MSPs Leads to More Errors and Downtime
John K. Waters: “Relying on AI for MSP services increases the risk of errors and system downtime due to lack of human oversight.”
Michael Perrinich: This one is a push. If you spot-check your technicians’ work, you spot-check AI output. That’s just quality control. AI can absolutely save you time and do real work, but 100 percent hands-off trust? No. The same oversight processes you’d apply to any team member apply here too. The risk is not AI. The risk is blind trust in AI.
Jasper Grewal: Computers do exactly what you tell them to do. The allure of AI is that it can go, “Wait, this doesn’t make sense,” and flag a problem. Could it lead to more errors? Yes, if you don’t respect that you get out of it what you put in. But that’s true of any system.
Andrew Moon: “Trust but verify” applies to people and computers alike. Especially with marketing content: if you just paste in whatever AI generates, you end up with generic output that doesn’t represent your business well. Review everything.
Claim 6: AI Integration Is Complex and Time-Consuming
John K. Waters: “Implementing AI into MSP operations is overly complex and will disrupt business for an extended period.”
Michael Perrinich: It can be, but not if you find the right partner. I became more interested in AI specifically because of how Syncro rolled their tools into the PSA directly. They put it in a place that was immediately useful for ticketing and smart search. When AI is delivered through a partner that provides the support and training, there is nothing complex about it at all. The complexity concern is real for MSPs trying to build something from scratch. But that’s not what most MSPs should be doing.
Jasper Grewal: If you identify what you want AI to accomplish and you’re using a tool like Syncro that has done the heavy lifting, the benefit is available from day one. If you’re building from the ground up, yes, it can be complex. Using Syncro’s native AI integration just empowers you without requiring you to become an AI architect.
Andrew Moon: I love seeing vendors solve this problem for MSPs. A lot of them are already removing the complexity that used to exist. But operationally, there is still work to do. SOP documentation is a great example. It’s one of the most time-consuming things MSPs put off. With AI, you can get 98 percent there in a fraction of the time. I literally generated a complete SOP for a sign company in a room with no domain knowledge about that industry. Walked through it out loud, typed the prompt, hit enter, and it produced a detailed SOP covering how to load the truck, what tools you need, and what the OSHA guidelines are. The people in the room were floored.
Jasper Grewal: The delivery layer matters too. Great if AI can extract and generate that SOP, but it has to be presented to the technician in a relevant, timely way. It needs to surface when and where it’s actually needed.
Andrew Moon: Exactly. The vision I’d love to see: dynamic SOPs that update as you debrief tickets over time, always current, automatically surfaced on the relevant ticket type. Any MSP that achieves that is not just more efficient. They are genuinely better.
Claim 7: AI-Driven MSP Services Are More Expensive
John K. Waters: “AI-driven MSP services are more expensive because of the advanced technologies involved.”
Michael Perrinich: Not anymore, generally. There may have been a time when it was. One area worth flagging: compliance. At August eTech, we work with HIPAA-regulated and FDA-regulated clients. We tell our team to use Microsoft tools because their data centers are HIPAA-compliant. Not because we’re putting client data into AI, but because we want to be in a compliant environment for everything we touch. Syncro is also HIPAA-compliant, which is one reason we’re building it into more of our workflow. But outside regulated environments, AI is just a feature now. It’s included in tools. It’s not even an upsell.
Jasper Grewal: I don’t see increased cost from the tool itself. The cost risk is misuse: a technician wasting time using the wrong AI tool in the wrong context, producing output that creates rework on the backend. In a regulated vertical like GCC High for 365, you pass the cost along. Your margin stays the same; the price reflects the compliance requirement.
Andrew Moon: AI has become power windows: it’s standard. The novelty premium is gone. Where you might still see cost is in the human element: setting up systems, governing usage, and verifying output. But you do that with every new tool. The difference is the scale of benefit once it is working.
Michael Perrinich: And someone needs to be watching the watchmen. As AI handles more of the research and documentation work, someone has to verify the outputs are accurate, the technicians are using the tools correctly, and the AI is not surfacing bad recommendations. That oversight is a real cost, but it’s a manageable one if you roll the tool out thoughtfully in stages.
Claim 8: AI-Powered MSPs Compromise Data Privacy and Security
John K. Waters: “AI systems in MSPs compromise sensitive data and increase cybersecurity risks.”
Michael Perrinich: I hate to say it, but this one is true, if it is managed badly. Let me give you a real example. I joined a Teams meeting and saw an unfamiliar label next to a participant’s name. I looked it up and a couple days later saw a TV commercial for the same product. It turned out that one of our people had signed up for an AI meeting recorder that joins your Teams calls and takes notes automatically. No installation needed. He added it to his schedule without asking for authorization.
I did what I do a lot: I read the terms of service. And there it was: “We don’t share your data, but if anyone acquires us, it’s up to them what they do with it.” For a company serving HIPAA-regulated clients, that is a serious compliance risk. I went to the C-level team, showed them the clause, pointed out that any meeting this tool attended could contain information that fell under HIPAA protection, and they immediately shut it down.
The lesson: read your privacy policies. Every single one. If you don’t know where your data is going, how it’s being stored, and what the terms are for future use, you are accepting liability you may not even be aware of. That is why we use HIPAA-compliant tools. Not because we’re putting client data directly into AI, but because we want our entire operational environment to fall within the right compliance boundaries.
Andrew Moon: This is one of the biggest opportunities for MSPs in the AI era. Clients trust us as their surgeons, not their janitors. We prevent problems; we don’t just clean them up. A two-to-three minute video or a short client communication explaining which AI tools are safe for their industry and which are not is a high-value deliverable. LinkedIn recently turned on an AI feature for all users by default without notifying them. That is another example of the kind of proactive education that MSPs can provide. You stay ahead of it, you flag it for your clients, and you protect them. That is exactly the kind of conversation that differentiates an MSP from a break-fix provider.
Jasper Grewal: Shadow IT carries the same risk whether AI is involved or not. Any unauthorized tool adoption can compromise data. AI just raises the stakes because of the volume and type of data these tools can access. Shadow IT could honestly be its own webinar.
Q&A: How Do You Measure ROI on AI Investments?
Jasper Grewal: Identify your objectives first. What are you hoping the AI tool will do for you? Where is the pain point? Measure what that problem costs you today, and measure whether the tool reduces that cost. One tool does something different from another, so start with the problem and work backwards to the measurement.
Andrew Moon: Time saved is the fastest ROI calculation. If your effective hourly rate as an MSP owner or operations lead is two to five hundred dollars per hour, saving even one hour per person per week creates a quick return. If I can save every person on my team an hour of low-value work, that’s real ROI.
Michael Perrinich: Think about every tool you’ve ever adopted in your stack. Why did you move from AV to EDR to MDR to XDR? Every upgrade required research, trial, testing, and measurement. AI is no different. It is not a magic bean. It is a tool. Evaluate it like one.
Q&A: How Can AI Improve Security for MSPs?
Jasper Grewal: The quick win is mail security. AI can evaluate the full context of a message and the intent behind it, rather than just flagging keywords. It can perform visual analysis on web pages and identify a spoofed Microsoft login page even when it’s hosted in Dropbox or Adobe. Contextual awareness of the totality of circumstances when making a security recommendation is the biggest benefit at the security layer.
Michael Perrinich: Microsoft does excellent work with Defender and Defender for Cloud. They stopped looking at keywords and started scanning entire file and message contexts. The AI is training on what phishing looks like at a contextual level, not just a word-match level. Teaching your users to tag suspicious emails correctly helps train the AI further. You’re not going to use ChatGPT to fix your cybersecurity posture, but embracing the AI components already in your security stack, and educating your users on how to interact with them, is where you see real security gains.
Q&A: How Should MSPs Evaluate Which AI Tools Are Right for Their Business?
Andrew Moon: Start with the tools your clients are already using. A lot of Gen Z employees are not buying into the full Microsoft stack the way MSPs who have been around for a while do. Tools like ClickUp and Notion are proliferating. The MSPs who test those tools themselves and form educated opinions about them can have a completely different sales conversation with their clients. I adopted Apple products because my clients were buying them, and using them daily made me better at supporting them. Same applies to AI tools outside the Microsoft stack.
Michael Perrinich: Every company selling an AI product wants to talk to you and give you a trial. Run them through your minimum standards checklist before you buy: SOC 2, ISO certifications, HIPAA compliance. Get a technical engineer on the call, not just a salesperson. A sales rep will say yes to everything; an engineer will tell you what is actually possible. And read the privacy policy. I cannot stress that enough. Read the privacy policy.
Jasper Grewal: Don’t buy it because it has “AI” in the marketing. Ask what specific problem this solves for your business. Where are you going to save time? Where are you going to multiply productivity? If you can’t answer that concretely, it’s not the right tool yet.
John K. Waters: Gentlemen, that is all the time we have. Thank you to Jasper Grewal, Andrew Moon, and Michael Perrinich for an outstanding conversation. And many thanks to Syncro for making this possible. Don’t forget: Syncro has three months free for today’s attendees when you sign up three-plus users on the annual team plan. Thanks, everybody.
See How Syncro Powers Your Business
Schedule a one-on-one walkthrough with a product expert to see the Syncro platform in action. No fluff — just a personalized look at how to unify endpoint management, service operations, and M365 workflows.
Frequently Asked Questions
No. All three panelists agreed that AI will not replace human technicians in any meaningful timeframe. The relationships MSP technicians build with their clients, including knowledge of individual business workflows, user preferences, and the kind of personal rapport that keeps clients retained for 20-plus years, cannot be replicated by any AI model regardless of training data. Where AI can help is in reducing the volume of low-level tickets so that technicians can spend more time on higher-value work, but the human touch remains the core of MSP service delivery.
AI is accessible and cost-effective for MSPs of any size. Tools like Microsoft Copilot are built into Windows 11 at no additional cost, ChatGPT is available for around twenty dollars a month, and PSA platforms like Syncro are integrating AI features directly into their existing toolsets. The panelists agreed that the “enterprise-only” framing is outdated, and that smaller MSPs may actually benefit more from AI leverage because they have fewer resources to throw at operational inefficiencies.
Yes, if AI tools are adopted without proper vetting. Michael Perrinich shared a real-world example where an employee added an AI meeting recorder to company Teams calls without authorization, and the tool’s terms of service allowed data to be transferred to any future acquirer of the company. For MSPs serving HIPAA, FDA, or PCI-regulated clients, this kind of unchecked AI adoption creates significant liability. The panelists recommended reading privacy policies carefully, preferring tools with clear compliance certifications (such as HIPAA-compliant data centers), and establishing internal policies on which AI tools employees are permitted to use.
AI is already embedded in many security tools MSPs use today. Microsoft Defender and Defender for Cloud use AI to scan entire messages and files for context rather than relying on simple keyword matching, which makes them significantly better at catching phishing attempts, malicious links, and fake login pages. Jasper Grewal noted that AI-powered mail security can perform “vision” analysis on web pages, identifying spoofed Microsoft login pages even when hosted on trusted platforms. The practical recommendation is to look at the AI features already built into the security tools in your stack and make sure you are using them fully.
Not when AI is built directly into the platform you already use. Michael Perrinich and Jasper Grewal both pointed to Syncro’s native AI integration as an example of AI that is accessible from day one with no additional setup, because the vendor has done the heavy lifting. The complexity concern is more relevant when MSPs try to build AI workflows from scratch or bolt together third-party tools. Andrew Moon added that SOPs are one area where AI dramatically reduces implementation complexity: prompting an AI to generate a detailed SOP in real time, covering tools, steps, and safety guidelines, is something that can happen in minutes rather than days.
Not meaningfully in most cases. The panelists agreed that AI has become a standard feature rather than a premium add-on, similar to how power windows are now included in every car. Most AI tools MSPs would use are either free, included in existing subscriptions, or available at low cost. The one exception is regulated environments: MSPs serving HIPAA, FDA, or government-regulated clients may need to pay for compliance-grade AI tools, such as Microsoft’s GCC High tier for 365, and those costs should be passed along as part of the vertical’s service pricing. The bigger cost risk is misuse, including wasted technician time from over-reliance on AI-generated content that requires significant cleanup.
The fastest ROI calculation is time saved. As Andrew Moon put it, if your effective hourly rate as an MSP owner or operations lead is several hundred dollars an hour, saving even one hour per week per person across the team adds up quickly. The panelists recommended starting by identifying a specific operational pain point, measuring how much time that problem currently costs, and evaluating whether an AI tool meaningfully reduces that cost. Michael Perrinich compared it to any other tool evaluation in an MSP stack: you research, trial, enable, and measure, and AI is no different in that respect.
Start with the problem you are trying to solve, not the tool. Jasper Grewal’s recommendation: before buying anything on the strength of an AI buzzword, ask what specific problem this tool is solving and whether you can measure the time or productivity savings. Michael Perrinich uses a minimum standards checklist for any new vendor, including SOC 2, ISO certifications, and HIPAA compliance status, and always requests a technical conversation with an engineer rather than just a salesperson. Andrew Moon suggested testing the tools your clients are already using, including platforms popular with younger employees who may not be in the Microsoft stack, so your MSP can support and guide adoption confidently.
Webinar Hosts
Jasper Grewal Business Leader, ROI Technology
Jasper Grewal is a business leader at ROI Technology, a cybersecurity-focused MSP serving SMBs in Washington state. With a career spanning desktop support, sysadmin, IT management, and DevOps, Jasper brings a practitioner’s perspective to AI adoption, focusing on where the technology adds genuine value versus where simpler automation already does the job. In this webinar, Jasper addressed AI’s role in security, technician relationships, and the limits of what AI can realistically replace.
Andrew Moon Business and Technology Strategist, Orange Nomad
Andrew Moon is a business and technology strategist at Orange Nomad who ran his own MSP until 2014 and has since coached nearly a thousand MSP owners on operations, sales, and business strategy. Andrew focuses on helping MSPs build “calm companies” by finding leverage through tools and processes, and he is an advocate for using AI to eliminate low-value work so MSP owners can focus on relationships and growth. In this webinar, Andrew made the case for AI as a sales and marketing force multiplier and shared how he uses it for content creation and SOP development.
Michael Perrinich Director of Operations, August eTech
Michael Perrinich is Director of Operations at August eTech, a full-service MSP based in central New Jersey specializing in biopharma, FDA, PCI, and compliance-regulated industries. With decades of IT experience and a strong focus on privacy and compliance, Michael brings a measured, process-driven view to AI adoption and has direct experience navigating HIPAA requirements when evaluating AI tools for MSP use. In this webinar, Michael shared real-world examples of AI-related compliance risks and explained how August eTech evaluates AI vendors using a minimum standards checklist.
Share
Related Resources
