Verdict

Okay, I need to get this off my chest. I've been running an MSP for a long time. I've survived the transition from break-fix to managed services, from on-prem Exchange to M365, from physical servers to cloud everything. I've watched vendors pivot their entire product lines twelve times in five years and I've adapted every single time. But AI? AI is doing something different to me. It's not making me angry in the normal way. It's making me angry in a complicated way. And I don't think I'm alone. Let me try to explain. The "AI-Powered" Label Is Meaningless Now and You Know It Somewhere around 18 months ago, every single vendor in our stack simultaneously discovered that their product had always, secretly, been powered by artificial intelligence. Your RMM? AI-powered. Your PSA? AI-enhanced. Your backup solution? Machine learning at its core, apparently - always has been, they just forgot to mention it for the past eight years. What does this actually mean in practice? It means there's a little sparkle icon somewhere in the UI now. It means the release notes say things like "leveraging advanced language models to surface actionable insights" when what it actually does is summarize a ticket into three bullet points that are somehow less useful than just reading the ticket. I'm not saying this technology isn't real or that it can't be valuable. I'm saying that vendors have absolutely weaponized the label to justify price increases, muddy the waters on what their product actually does, and make you feel like you're falling behind if you don't adopt it immediately. The signal-to-noise ratio is basically zero right now. Our Clients Think We're Obsolete (Thanks, ChatGPT) This one stings a little, I'll be honest. How many of you have had a client - a client you've supported for years, a client whose server room you've physically crawled into at 11pm in January - casually mention that they "just asked ChatGPT" and got an answer? How many of you have had someone say "can't AI just handle that?" when you quote them for a project? There's a fundamental misunderstanding happening at the client level about what AI can and cannot do. The same person who will forward a phishing email to every contact in their address book is now completely confident that a chatbot can replace their IT support. They watched one YouTube video about AI agents and now they think the whole MSP model is a relic. What they don't see - what they will never see until something breaks catastrophically - is the decade of institutional knowledge you're carrying around. They don't see that you knew their ancient line-of-business application would choke on a Windows 11 upgrade because you were in that exact situation three years ago with a different client. They don't see that you caught the anomalous login attempt at 2am because you actually understand their environment. ChatGPT doesn't know their environment. It doesn't know anything about their environment. It hallucinated a fix for their specific Brother printer driver issue and they followed it and now something is worse. But try explaining that without sounding defensive. The Internal Pressure to "Have an AI Strategy" Raise your hand if someone - a vendor, a peer, a conference speaker, a well-meaning industry analyst - has told you that you need an AI strategy. Now raise your hand if any of those people clearly defined what that actually means for a 3-person MSP serving 30 SMB clients in a regional market. Yeah. There's enormous social pressure right now, especially in MSP communities, to be "doing something with AI." The thought leadership content is relentless. The webinars are endless. Every conference session is either directly about AI or has AI shoehorned into it because that's what gets butts in seats. And all of it is oriented toward the idea that if you're not building AI into your practice RIGHT NOW, you're going to be left behind. Some of that is legitimate. Some of it is vendors trying to get you to buy new tiers of products. Most of it is just noise generated by people who are also confused but have a platform and a deadline. The honest truth is that most of us are already using AI - just not in a way that makes for good LinkedIn content. We're using Copilot to draft client emails faster. We're using Claude or ChatGPT to help write PowerShell scripts we'd otherwise spend 45 minutes on. We're using AI-assisted tools in our security stack that have been there quietly for two years. We just don't call it an "AI strategy." We call it "getting work done." Security Is Now a Nightmare in New and Exciting Ways I have a whole category of gray hair that is specifically AI-related security gray hair. The phishing emails are better. Dramatically better. The era of catching a phish because it said "Dear Valued Costumer" is largely over. We're now dealing with grammatically flawless, contextually plausible, sometimes eerily personalized social engineering attempts. Voice cloning. Deepfake video for executive impersonation. AI-generated fake invoices that pass a casual visual inspection. And what's my client base? Small and medium businesses with no dedicated security staff, limited budgets, and a general philosophy of "it won't happen to us." I am fighting a war with increasingly sophisticated weapons on the adversary's side, and my clients are mad that I want to add MFA to their email. Meanwhile, I'm supposed to explain to them that the voicemail from "their bank" was AI-generated, and also yes you do need that security awareness training, and no I can't just make it so this doesn't happen without you participating in basic hygiene, and by the way your cyber insurance carrier wants you to fill out a 40-page questionnaire. It's exhausting. The Automation Anxiety Is Real, Even If We Don't Talk About It Here's the one nobody wants to say out loud in MSP circles: a non-trivial number of us are at least a little worried about what AI-driven automation actually means for our business model long-term. Our value has always been expertise plus labor. We know things, and we do things, and clients pay us because knowing and doing those things has historically required skilled humans. AI is, in a very real and accelerating way, compressing the labor component. Ticket summarization, documentation generation, basic troubleshooting flows, first-level triage - these are things that AI is genuinely getting good at. The optimist answer is that this frees us up to do higher-value work. And I believe that, mostly. But it requires acknowledging that some of the billable work we do today - some of the work that's currently embedded in our contracts and our per-seat rates - is going to get automated. The question isn't if, it's when, and whether we're the ones doing the automating or whether we're the ones being automated around. The MSPs who are going to win are the ones who figure out how to use AI to deliver more value at the same price point, rather than waiting for clients or competitors to force the conversation. But that requires investment, experimentation, and frankly some tolerance for things not working - and most of us are already running lean. What I Actually Think (and What I'm Actually Doing) Look, I'm not a luddite. I've been in this industry long enough to know that every wave of "this changes everything" both does and doesn't, and the survivors are the ones who find the practical path through. AI is genuinely useful to me right now in specific, concrete ways. Script generation and review. Documentation drafting. Summarizing long email threads before I respond to them. Helping me think through architecture decisions by acting as a sounding board. These are real time savings that I've folded into my actual workflow. What I've stopped doing is reading every breathless "AI will transform your MSP" article as if it contains actionable intelligence for my specific situation. Most of it doesn't. Most of it is written f