Practical Perspective on the Risks, Trends and Decisions Shaping Security
Issue No. 1 · October 2026
A letter from Tim Burke, President & CEO, Quest Technology Management
In this Issue
1. What we missed
2. Why I started the Ledger
3. Ransomware in the middle of the market
4. When your provider gets bought
5. The big vendors are coming down-market, and they’re bringing AI
6. The boring stuff
7. When the vendors change the rules
8. What we’re telling clients about AI
9. What we’re watching next
To our clients, partners and friends:
If you run technology for a company with a few hundred employees and asked me over coffee how things are really going, here’s what I’d tell you. There’s a lot of anxiety and confusion. And one question I hear more than any other: where should we spend our budget?
Most of the leaders I talk with aren’t short on effort or good intentions. What they’re short on is certainty. Do I have the right people? Was the tool we bought turned on? Are we as protected as I think we are? In our experience, the honest answer is often “not quite.”
Here’s something we see again and again. In many of the incident responses Quest has handled recently, the door the attackers walked through was a baseline patch that everyone believed was being applied but wasn’t. Not a nation-state. Not some exotic AI attack. A routine update nobody applied.
When I shared that recently with a CIO who had no connection to any of those incidents, he didn’t say, “We should check.” He said it couldn’t be. I understand the reaction. Nobody likes hearing it. But the gap between how we think our environments work and how they actually work is where most of the trouble in this business lives.
This is the first issue of The Mid-Market Security Ledger. Each quarter, I’ll write down what we’re seeing, what we think it means, and, just as important, what we don’t know yet.
What we missed
Even good teams miss things. It happens to the best of us, and our industry is no exception. Here are a few worth sharing.
The first is a habit our whole industry shares, Quest included. For as long as I’ve been in this business, buyers, vendors, and resellers have chased the magic bullet. Buy the right product, and all’s good. Those products are often very good. But a good product doesn’t fix a process that nobody is running.
I know of a company that spent roughly $2.5 million on new security tools in a year when, as far as I could tell, it still wasn’t patching consistently. Everybody in that deal was happy. The IT team got its new tools. The vendor got paid. The reseller got paid. The only party that didn’t win was the company. We all enable each other, and sometimes the reseller is us.
The newest version of that habit is AI. A year ago, the pitch was simple: bring it in, people start using it, and you save a ton of money. What we underestimated was what happens to your data once dozens of AI-enabled tools start creeping around inside it. Marketing tools, sales tools, things nobody thought of as AI tools at all. We still don’t fully know what they’re doing with that data. The bigger problem is that we acted as if we did. We want certainty, and certainty doesn’t really exist.
The last one is personal. It took me years in this business to fully appreciate the difference between an operational business and a financial business. I do now, and it matters to you more than you might think. More on that below. Back to top
Why I started the Ledger
This is the first issue, so it’s fair to ask why it landed in your inbox. Every week, my team and I sit down with technology leaders who are working through the same questions you are. Across those conversations, we see patterns that are hard to spot from inside any one company. The Ledger is my way of sharing them: what we’re seeing, where we think things are headed, and what’s new at Quest.
It isn’t a sales pitch. Think of the magnet a good plumber leaves on your refrigerator. You don’t give it a second thought until the sink backs up, and then you’re glad it’s there. I’d like this letter to work the same way. If something in it sounds familiar, you’ll know who to call.
You can expect the Ledger every quarter.
Ransomware in the middle of the market
Black Kite found that mid-market firms accounted for 73% of ransomware incidents in 2023. That matches what we see, with one caveat: nobody agrees on exactly what “mid-market” means. Call it a company big enough to have something worth stealing and small enough to be stretched thin.
The sentence I hear most from leaders at those companies is “Why would anybody target me?” I can’t tell you how many times we hear it. The honest answer is that nobody is targeting you personally. They’re pushing on doors.
The business model isn’t complicated. Picture someone walking down a street trying door handles. Most are locked, so he moves on. He isn’t going to spend weeks breaking into Fort Knox. When a door opens, he slips in, checks for cameras, finds the file room, and takes whatever he can sell. AI hasn’t changed that model. It just lets him try a lot more handles, a lot faster.
Why are so many doors open in the middle market? Attackers don’t take nights or weekends off. Plenty of mid-market companies aren’t watching around the clock. Some aren’t even watching closely during business hours. They bought tools that may or may not be installed properly, and they may or may not have the people to run them.
In most of these companies, leadership expects IT to be handled in-house, often by four capable, well-meaning people who are asked to do everything. Hiring a dedicated security leader can cost a quarter of a million dollars a year, and most companies this size can’t justify it. The result is enterprise-grade problems with small-business staffing, and attackers know it. They go where the defenses are thinnest.
What changes the math isn’t a giant overhaul or another product. It starts with three questions. What do we need to protect? What protects it today? And, critically, is that protection working? I don’t mean a list of what you bought. I mean: Is it deployed? Is it operational? Who is monitoring it, and what happens when it raises an alert?
In my experience, maybe 20 items on a checklist, verified every month or quarter, give you real confidence that when someone pushes on your door, it holds. You won’t be Fort Knox. You don’t need to be. You
need to be the house they skip. Back to top
When your provider gets bought
Our industry is consolidating fast. Investment firms and roll-ups are buying managed service providers at a steady clip. Every press release says nothing will change for customers. In our experience, that promise has a shelf life of about two quarters, sometimes less.
Here’s why. An acquisition is a financial transaction, not an operational one. Someone paid for that company, and the purchase has to pay for itself. In a services business, the highest costs are people, services, and tools. That’s where the savings come from.
The six- or ten-person shop that was at your beck and call gets folded into shared security and network operations centers. The people you relied on may no longer be there. Nobody did anything wrong. The new owner is simply playing a financial game with different rules. If you don’t know the rules, you can be really challenged.
The sharpest surprise is usually in the contract. We’re regularly brought in after a customer discovers that what they thought their provider did doesn’t match their service level agreement (the contract that spells out what you’re owed). “Bob used to come right over and handle it.” That’s nice, but it isn’t in the agreement. The old owner let Bob do it anyway. The new owner goes by the contract. Back to top
The big vendors are coming down-market, and they’re bringing AI
Large technology companies are pushing AI into every corner of the market, including through providers like us. Some of that is genuinely good news. Good tools are getting cheaper and closer to you.
It helps to remember what’s driving it. Search had to figure out how to make money, and now AI does too. The companies building it are working very diligently on exactly that. What they’re delivering is a product. What they aren’t necessarily delivering is the oversight that goes with it: which files the tool reads, what it searches for, and where your data goes next.
The result is that many mid-market companies are now exposed to AI agents they may not know they have. At Quest, we keep an internal list of agents we’ve blocked. If someone on our team tries to log in to one, they simply won’t get there. That’s not because those products are bad or proven unsafe. It’s because we don’t know what they’re doing yet, and that’s reason enough.
The new assistant that promises to book your flights and run your life may turn out to be wonderful. Be careful anyway.
This will hit the middle of the market hardest, because these companies rarely have the audit, compliance, and validation resources to keep track of who is using what. A tool you didn’t choose, configured by someone you’ve never met, is only as good as the monitoring and response behind it. Anyone can read an alert at 2 a.m. What matters is whether someone acts on it. Back to top
The boring stuff
Nobody has ever been promoted for patching on time, and plenty of people have been fired for not doing it. That imbalance explains more breaches than any headline threat.
This past quarter, we saw several companies compromised through a known flaw in their firewalls. The vendor had already released the fix. When we asked whether they’d seen it, the answer was some version of “Yes, but we see so many patches.” It’s like hearing the hurricane warning and staying put because there’s always a hurricane warning.
The most dangerous gap I see isn’t technical. It’s at the top. Executives spend real time understanding balance sheets, cash flow, and production. Technology is now just as central to whether they’re still in business two weeks from now, and most leadership teams haven’t caught up. They hand it to IT and accept the answer.
Plenty of executives couldn’t tell you whether their IT team truly has things under control. That’s not a knock on IT. Most of the IT people I meet are knowledgeable and well-meaning. The problem is that nobody validates. Ask your IT lead, “Are we all patched up?” and you’ll probably hear, “Yeah, we’re fine.” He may even believe it. But unless someone checks, he doesn’t really know, and neither do you.
One more, and I’m still shaking my head. At one company, the finance team had disabled multi-factor authentication (the second step that confirms it’s really you logging in) because it was too much of a bother. That’s exactly how the attackers got in. About $2 million was wired out of the company’s bank account to an overseas account. We caught it in time and recovered the money. As far as I can tell, nothing changed afterward.
None of this calls for a magic bullet. Most of security isn’t difficult, and you already know what to do.
AI hasn’t changed that. It runs the old mistakes faster, at a scale no human team can match. An unpatched system was always a risk. What changed is how quickly somebody finds it. The fundamentals didn’t get less important this year. They got less forgiving. Back to top
When the vendors change the rules
A licensing change is a price increase that doesn’t want to be called one. Two are on my mind this quarter.
The first is the move to paying for AI by usage. Many AI tools started out as a flat monthly subscription per user. More are shifting to pricing based on tokens, the small chunks of text an AI model reads and writes. Ask 99% of the people in your company how many tokens their last question cost, and they’ll have no idea. “Here are 100 free tokens” doesn’t mean anything to a business user. For the companies selling AI, this is the gift that keeps on giving. For a mid-market budget, it’s a cost that’s hard to see until the invoice shows up.
The second is plain complexity. Productivity and cloud licensing now come in so many tiers and add-ons that even capable IT teams struggle to match the right license to the right person. We’ve seen companies decide to just give everyone the top tier. It’s simple, and it’s expensive. In one case, roughly 80% of what each user was paying for wasn’t being used. Who actually needs the top tier? Who needs the middle one? Who needs phone calling built in? Those are answerable questions, but somebody must ask them.
What does a sensible person do about all this? Treat licensing like any other recurring cost. Review it at least once a year. Ask what’s actually being used. And before anyone signs up for usage-based AI pricing, ask a simple question: who is watching the meter? Back to top
What we’re telling clients about AI
Here’s what we know for sure about AI and security: the attackers got faster, the defenders got louder, and most of the advice you’re getting comes from someone with something to sell.
What I’m telling clients mostly comes down to two words: slow down. It’s okay. I talk with a lot of CEOs and CFOs who are convinced that if they don’t have AI fully operational in the next two months, they’ll be out of business. They won’t. They just won’t.
I wrote recently that while the big AI labs talk about pacing themselves, your company should do the same. The questions that matter haven’t changed. Who and what has access to your systems? Which data is sensitive, and is it labeled in a way a machine can respect? If something goes wrong, how would you find out, and how fast?
I’d also tell you that the gloom-and-doom story about AI’s cost and power isn’t the whole story. Yes, enormous sums are going into data centers and chips. I suspect some of that capacity will look overbuilt in two or three years. We’ve seen this before. Fiber optic networks were overbuilt in the late 1990s and early 2000s, and it took nearly two decades to use what was laid.
Meanwhile, engineers are designing AI that needs far less computing power, and companies are bringing it in-house. Quest runs its own AI platform on its own hardware, isolated from the outside world. Today you can run a capable AI model on a well-equipped desktop computer. Over time, that trend should make AI more affordable and easier to control for mid-market companies.
Now for what I don’t know. I don’t fully understand where AI systems that work with each other, create things that didn’t exist before, and find their way around controls will end up. Some industry leaders have argued recently that we still control what these systems can and can’t do, and that a problematic capability can simply be turned off. My instinct is that they’re right, and the genie hasn’t left the bottle. But I don’t know. I’d be suspicious of anyone who tells you they do. Back to top
What we’re watching next
We don’t make predictions, but we do keep a list. Here’s what’s on it between now and the first quarter of 2027.
At the top is the temperature of the AI debate. Whatever any of us believes about AI and the end of civilization, what matters for business is what the hysteria does. Lately, statements from industry leaders and agreement at the highest levels on what AI companies need to do seem to have restored some confidence that people are still in charge. I’m watching whether that holds.
Right behind it is the AI economy itself. A large share of recent U.S. economic growth is tied to AI spending, and several big public offerings are on the way. Is that a sturdy economy or a fragile one? I don’t know. Neither does anyone else, which is exactly why it’s on the list.
Then there’s fatigue. People eventually get tired of fear, uncertainty, and doubt, and they throw up their hands. When leadership teams get worn out, they postpone security decisions, and attackers find open doors in postponed decisions.
Finally, I’m keeping an eye on the meter: how usage-based AI pricing shows up in mid-market budgets over the next two quarters. My guess is that a lot of companies will learn what a token costs the same way they learn most things about licensing, when the invoice arrives.
In January, we’ll look back at the fourth quarter of 2026 and ahead to 2027, including how the items on this list played out. Back to top
An open door
If any of this sounds familiar, I’d love to hear your version. Our doors are open, and the coffee is better than you’d expect. You don’t need a problem to reach out. But if you find one, we’ll ask the same question we’ve been asking since 1982: How can we help?
Tim Burke, President & CEO, Quest Technology Management
