Lately we have been spending a lot of time with customers revisiting what I would call the cybersecurity essentials. These organizations face the same risks as companies ten or a hundred times their size, because they use the same technologies. What they typically don't have is the budget or manpower to protect themselves the way a large enterprise can.
Revisiting these essentials, it's clear that the list itself hasn't changed. You still need endpoint protection. You still need a firewall. You still need email security, because email is still a primary attack vector. And you still need to protect identity, starting with multifactor authentication.
What has changed is the urgency of your response to threats.
We have entered an era where risks that sat dormant inside environments for years are now being discovered, and quickly. Earlier this year, a researcher using an AI coding tool found a remotely exploitable flaw in the Linux kernel's NFS file-sharing code that had been there since 2003. That's 23 years, and no one had caught it until AI was put to the problem.
It doesn't matter whether the good guys or the bad guys find a flaw like that first. Once it's out, everyone is scrambling. And many of these flaws sit on services at the edge of the network, exposed directly to the internet, rather than behind the walls where we are used to patching on our own schedule.
The pace is the real story. New threats used to emerge monthly or weekly. Now they appear daily, sometimes hourly, because attackers are feeding everything they can into AI and acting on whatever comes back out.
Think about what that means for a single attacker. One person using AI can behave like a thousand attackers, or ten thousand, probing your environment in many different ways at once. It costs them nothing more to attack a 50-person company than it does to attack anyone else. If you have an unpatched vulnerability facing the internet, you are at risk, regardless of your size.
For years the challenge was visibility. Every tool fed alerts to a person, and someone had to decide whether each one was a real risk. AI has helped enormously here. It can sort through that volume and surface what matters, which cuts down on how much a human has to look at.
But once AI flags a problem, how long does it take to respond, mitigate and fix it? That is where most organizations are now falling behind.
Patching is a good example. The traditional approach was to roll a new patch out to a pilot group, make sure it didn't break anything, and then slowly roll it out to everyone else. We can't wait that long anymore.
That means making some decisions ahead of time, based on how your business runs. If you operate a pipeline, you can't just shut it down. But if you offer guest wireless at a retail store, you can turn it off the moment a threat appears, because it isn't critical to the business. Once you have decided what can be shut down and what can't, you can give your tools those parameters and let them act. We have to turn over some of the reins, because as human beings we are simply too slow to keep up.
In practice, I see two ways to stay ahead of this.
The first is to build a patch management and vulnerability program that is tied directly to your security tools, so that fixes and protective changes happen automatically and quickly.
The second is to design your environment so that very little of it touches the internet in the first place. Anything that isn't essential shouldn't be exposed. That is a long-standing principle, but it isn't always put into practice. When someone tells me they need a web server, I'll ask whether they really do. Sometimes the better answer is to redesign the architecture so they don't.
This applies to cloud-first companies, too. An organization with no servers on-premises still has users with laptops, and those laptops are now the boundary of the organization. An employee working from a coffee shop is just another IP address on the internet, waiting to be probed. We can't push endpoint patches off until later anymore.
Enterprises generally have the people to work through all of this. Small and midsize businesses usually don't. Realistically, they have two choices: build an enterprise-like security team in-house, which can cost hundreds of thousands of dollars, or hand this part of the business to a managed service that runs it for them at scale.
If there is one area where all of this comes together, it is identity. Attackers don't hack in anymore. They log in.
The techniques they use to get credentials will keep changing, whether that's phishing, voice phishing or a deepfake. The goal never changes. They want the identity, the credentials or the token, because once they are logged in they can work on staying there.
This is where AI-driven analytics make the biggest difference. Say I am logged in from one IP address, and two seconds later I appear somewhere else, on a different computer and a different browser. That's suspicious, and the right response is to lock the account immediately, not minutes or hours later. Stopping identity-based attacks the moment they happen cuts down the dwell time an attacker has inside your environment.
A fair question I hear is whether bringing AI-powered tools into an environment puts the organization's own data at risk.
The first step is explaining what is actually being collected. For the most part, we are looking at event logs and transactions, not the content people create. Security tools do increasingly inspect encrypted traffic, because that is where attackers know they can hide. But what feeds the AI analysis is metadata: who, from where, which IP address, which identity. The security tools themselves make the determination about individual files and transactions.
That combination is what lets us spot behavior that no single tool would flag. A security tool won't call it malicious if someone moves files to Google Drive. But if a company doesn't use Google Drive and one person suddenly moves a thousand files there, that's worth asking about.
The second step is transparency. At Prescriptive we will, for example, show customers the tools we use, the data controls in place and what happens at every step. If someone is still uncomfortable with AI, I might take the conversation back to the fundamentals: understanding risk. Attackers are already using AI against you. Without AI-backed tools on your side, you will always be behind the curve, even with the best tools in the world.
None of this means a small business needs every tool a multibillion-dollar enterprise has. The risk is the same at any size, but not necessarily the level of exposure: the services you run, how many people have access, and how critical each system is to the business.
That's why the work starts with understanding your business and how it uses technology. Once we understand both, we can show you where you are exposed and lay out the options. The decision on what to implement, based on cost and impact, is always yours. No one expects a business owner to be a security expert, but you understand your business better than anyone.
At Prescriptive, we take what we call a customer zero approach. We are a midsize business ourselves, and we protect our own environment with the same tools we use for our managed services customers. When a tool, policy or procedure stops meeting the standard, we replace or reconfigure it for ourselves first, and then roll the same change out to our customers. We aren't beholden to any one platform.
There is good news here, too. Most major security vendors are building AI analytics and response into their products, often without making them more expensive. Delivered as a managed service, that puts enterprise-level protection within reach of organizations that aren't enterprises.
What was standard a year ago is still on the list. It just isn't enough on its own. The question for every small and midsize business now is whether its essential tools are backed by AI and set up to respond at the speed attackers move. If they aren't, they are probably too slow.
Prescriptive can help you understand where your business is exposed, which protections fit how you operate, and how to respond at the speed today's threats demand. If you're ready to take a fresh look at your essentials, contact us. We'd welcome the opportunity to help.