Key Highlights
- Agentic AI = digital workers with decision-making autonomy: Unlike rigid "if-this-then-that" software, AI agents combine knowledge, tools, and generative AI to autonomously figure out how to complete tasks like monitoring emails, creating service tickets, troubleshooting with customers, and escalating internally without human intervention.
- The competitive gap is already forming: Integrators are already using AI agents for service tickets and tier-one support, scaling operations without proportional staff increases.
- Service differentiation can finally be codified and scaled: AI agents automate an integrator's specific service approach (not generic workflows), scaling what makes their business unique across every customer interaction.
- The investment decision is simpler than you think: Don't try building the perfect assistant; pick one narrow use-case (service acknowledgment, password resets, firmware checks) that's causing daily friction, start there, and expand as you learn.
This article originally appeared as the cover story in the November 2025 issue of Security Business magazine. Don’t forget to mention Security Business magazine on LinkedIn or our other social handles if you share it.
You’ve probably heard the term thrown around at industry conferences and in vendor pitches: Agentic AI. But if you’re like most security integrators and consultants, you might be wondering what it actually means, and more importantly, whether it’s something you need to pay attention to or just another in a sea of tech buzzwords that will eventually fade away.
The short answer? This one looks like it matters. Agentic AI represents a fundamental shift in how artificial intelligence can work for a security business, moving beyond simple chatbots and analytics to create digital workers that can actually take action and make decisions within defined boundaries – to solve many of the mundane and repetitive tasks that tend to cut into an employee’s average day-to-day duties.
To cut through the hype and get to what security executives really need to understand, I spoke with Don Morron, founder of HighlandTech, an AI automation agency that builds custom AI agents for small and medium-sized businesses. With 13 years in the security industry, Morron brings both technical expertise and practical insight into how integrators can actually use this technology.
I also gathered perspectives from Steve Lindsey of LiveView Technologies, who recently spoke about agentic AI’s transformative potential at the Securing New Ground Conference, hosted in New York by the Security Industry Association (SIA) in October.
What Exactly is Agentic AI?
Morron breaks it down simply: “Agentic actually is a word to describe agents,” he explains. “To define agents, think of them as digital workers that will automate some, most, or all of a task – usually some business process.”
Morron puts it in simple terms: Traditional software follows rigid instructions – ie. if this happens, do that. This method is predictable and reliable, but inflexible. AI agents, on the other hand, have what Morron calls “agency” – a degree of autonomous decision-making that allows them to figure out how to accomplish a task, not just execute pre-programmed steps.
“Agency is free will, like we have as people,” Morron explains. “As a person, you have free will to decide what you need to do, based on some knowledge or tools, to execute on some task. The concept is the same for AI agents.”
Under the hood, AI agents combine three elements: knowledge (documents, procedures, product data), tools (business systems like email, CRMs, or service management software), and that crucial element, agency – which is powered by generative AI, the same technology behind ChatGPT.
Why it's Different From the AI You Know
Integrator and consultant executives are already familiar with AI. The impact of video analytics, facial recognition, object detection, and more has been felt within physical security for years. What makes agentic AI different?
“AI, in general, has been around for decades, and computer vision is what this industry has lived on,” Lindsey explains. “Some of those breakthroughs – like when we went from just motion detection into object detection and classification about 10 years ago – really revolutionized the way we could do proactive video monitoring.”
But here’s the key distinction: “What’s interesting about agentic is when you think about computer vision, it is a very static and algorithmic training of a model,” Lindsey says. “With agentic, we think about things like visual video language models and large language models, which is a totally different approach to solving the problem.”
Traditional computer vision looks for specific things: a person, a vehicle, a package. Agentic AI can understand context and behavior. “The ability for a camera to transform what’s happening in the physical space into data unlocks a whole bunch of capabilities that are desperately needed in this industry,” Lindsey says. “[For example], the need to pull red actors out of a sea of green actors – the only way you can do that is to look at behavior in context, and you can’t do it with simple object classification.”
What Most People Get Wrong About Agentic AI
Before diving into practical applications of agentic AI for a security business, it is worth clearing up some common misconceptions – because if you have experimented with AI tools and been disappointed, you’re not alone.
It doesn’t work: “Immediately what comes to mind is that, ‘Oh, I used it and it doesn’t work,’ or, ‘I used it and it made up some stuff,’” Morron says when asked about the biggest misconceptions he encounters.
The problem, he argues, isn’t usually the AI technology; it’s the approach. “I think the misconception comes back into, ‘Well, it didn’t work the way they said it was going to work, because I prompted it and it gave me something weird,’ and then they immediately give up," he adds. “They gave it the most basic instructions ever, thinking they were going to get something good.”
Consumer AI tools like ChatGPT are general-purpose assistants. They are impressive, but they are not trained on your business, your processes, or your products. Building effective AI agents requires proper configuration and training on a specific use-case. Prompts are important, but building an agentic system takes much more.
Hallucinations: You’ve probably heard about AI “hallucinations” – when AI generates information that sounds plausible but is completely fabricated. That is a real issue, and Morron doesn’t sugarcoat it: “Hallucinations are a real problem within the AI agent development space, and even the AI systems space,” he says.
The solution? Proper design. Morron emphasizes starting with what he calls “utility agents,” which are AI agents with very narrow, well-defined tasks. “We want it to have a narrow task so it doesn’t make stuff up and we can trust its level of agency,” he explains.
Promises vs. reality: There’s a gap between what vendors promise and what’s actually practical today. “Thoroughbred agents, the way Silicon Valley looks at them, are given a goal, and it accesses all your tools and data to figure it out on its own,” Morron says. “You have [Silicon Valley] companies selling these agents, but they don’t really work that well, and so there’s a lot of over-promise, under-deliver.”
The takeaway from the misconceptions is clear: Deployment of AI agents for your business should be focused on practical, narrow applications rather than waiting for the sci-fi version of AI agents to become reality.
From Theory to Practice
Morron founded his company, HighlandTech, specifically to make AI automation accessible to small and medium-sized security businesses – companies that don’t have in-house software development teams or million-dollar IT budgets. “We do that by building custom AI agents,” he says.
With more than a decade in the security industry, he understands the unique challenges that small and medium-sized integration companies face. “These are a lot of blue-collar guys who might have started as technicians and have now started a company,” Morron explains. “They run a great business and they really care about their customers, but as they grow, they don’t always [evolve] their business processes.”
HighlandTech helps integrators break out of that pattern by codifying better processes into AI agents – essentially using technology to force process improvement while simultaneously automating the work. “I’m passionate about being able to remove the things that are repetitive and mundane, so we can focus on what makes humans great – relationships, empathy, and conversations like this,” he says.
Since launching his company, Morron has created many AI agents specifically for security integration companies. Here are a couple of ways they are using agentic AI right now to transform their operations:
Automating Service Requests: Morron is currently working with a client on one of the more practical applications for a security integration company: automating the service ticket process. The agent monitors email inboxes, creates service tickets in whatever software the company uses, such as SedonaOffice, acknowledges the customer, and handles internal escalations.
The technical process is straightforward: The AI reads the email as it comes in, creates a SedonaOffice service ticket, and replies to the email acknowledging receipt of the request with details about the service ticket and a promise to contact them as soon as possible. Meanwhile, in the background, the AI agent is escalating the ticket internally every hour until it is addressed by a human. The customer gets immediate acknowledgment – which is often what they want most – while it helps the integrator’s service team get organized with tracking and automatic internal reminders. Nothing falls through the cracks.
What makes it powerful is the agent’s ability to embody the specific integrator’s approach to service. “The idea is to define what ‘service’ means to that integrator,” Morron says, “because the only difference between one integrator and the next, in my opinion, is their ability to serve the customer.”
It is worth pondering: If most integrators sell the same products and use similar installation techniques, service is often the only true differentiator. The ability to codify a company’s specific service approach into an AI agent means you are automating while also scaling the uniqueness of a business.
Tier-One Technical Support: The next level builds on basic ticket creation by having the agent actually troubleshoot with the customer. Morron has set up AI agents to provide tier-one tech support. He notes that this involves the very basics, and the agent will interact with the customer to ask things like “Is the firmware updated?” or “Is it plugged in?”
But the agent can be programmed to delve deeper than a basic checklist. Here’s a vendor-agnostic example scenario Morron describes: “Let’s say they contact an integrator and say their key card system is not working. The agent could come back and try to triage, maybe through a couple of emails – ‘I’m sorry your key card system’s not working. Can you please tell me what make and model you’re working with, and can you please specify the issue, and what not working means?’ Once the customer responds, the agent could say: ‘Here are some self-help tips that you can do yourself while we get someone to help with a service call.’”
This accomplishes multiple objectives. “Now the customer can deal with internal issues like a password that needs to be reset, and they’ve been immediately acknowledged and they feel heard – which is what they really want,” Morron says.
Agentic AI should quell the common fear that AI is going to take a human’s job. Instead, it’s about scaling capability. An integrator experiencing 30% growth will not need to add 30% more support staff if AI agents handle the increased volume of routine requests.
The business impact is obvious: Fewer truck rolls for issues customers can solve themselves, faster resolution times, 24/7 support availability, and technicians spending their valuable time solving problems that actually require their expertise.
Lindsey frames the broader opportunity in terms that should resonate with any security executive managing budget constraints: “This industry is ripe for being able to use the force multiplying benefits of agentic AI,” he says. “This is where the agentic AI’s reasoning model can start handling lower-priority requests and automate a lot of that, and then allow the humans to focus on the high-priority issues that are coming in.”
This should quell the common fear that AI is going to take a human’s job. Instead, it’s about scaling capability. An integrator experiencing 30% growth will not need to add 30% more support staff if AI agents handle the increased volume of routine requests.
Voice AI: The Next Frontier
While email-based agents are practical today, voice AI represents the next evolution, and it’s closer than you might think. “Rather than an email, we could [address these service requests] with a conversation,” Morron says. “[An AI agent] can talk with the client to walk them through [simple tasks]. I have a demo right now where I can do that.”
He adds: “There’s a saying in the AI community – this is the worst it’ll ever be. With the accelerated pace of growth, there’s no question in my mind that a year from now, voice AI is probably going to be indistinguishable from a human voice.”
For now, there are still challenges, like response latency, natural tone, and the handling of different accents; however, these are technical problems being actively solved, not fundamental limitations.
This leads to another security application that many integrators with monitoring operations are likely thinking about, but are hesitant to voice: Can AI agents replace monitoring center operators?
The honest answer? It’s already happening in limited ways. “When we look at alarm monitoring, they are already doing it, to a degree, by filtering out nuisance alarms,” Morron says. “That’s agentic behavior – it has some agency to say, ‘This is a nuisance alarm and I’m not going to put it in front of an operator.’”
But full replacement of human operators raises both regulatory and practical roadblocks. “Just because AI can perform some of the tasks in a monitoring center doesn’t mean you should,” Morron says. “These are monitoring centers, and there’s got to be some level of certification and checks and balances.”
The practical path forward is likely a tiered implementation, with AI agents handling and filtering routine alarms, with human operators focused on genuine emergencies and complex situations. This maximizes efficiency while maintaining appropriate human oversight.
Technical Hurdles
Starting with narrow-task AI agents makes sense in the face of the technical challenges that still need to be solved by Silicon Valley. “The challenge we have today around AI agents is mostly around memory – what they call persistent memory,” Morron explains.
An agent that creates service tickets doesn’t need to remember a customer’s full history; it just needs to handle the immediate request. These memory requirements are contained and manageable. To go beyond that will take much more dedicated memory, which Morron says is not inexpensive.
Another challenge specific to the security industry is the proprietary nature of many products and manufacturers’ desire for secrecy. “One problem with building an omniscient service agent that knows all about products is that our industry doesn’t do a good job of making those products accessible publicly,” Morron says.
This is both a challenge and an opportunity. For integrators, focusing agents on the specific product lines they support and organizing that documentation can create a competitive advantage. For manufacturers, making product information more AI-accessible could become a significant selling point.
The Bottom Line: Should You Make the Investment?
Agentic AI isn’t science fiction, and it’s not just hype. It is a practical technology that security integrators are already using today to scale their service capabilities, respond to customers faster, and free up technical talent for higher-value work.
As AI agent solutions become more available, security integrator executives face an important strategic decision: Purchase off-the-shelf solutions or invest in custom development. Morron says it comes down to how much a specific process matters to an integrator’s competitive advantage.
For standard business processes such as basic scheduling, invoice processing, and standard reporting, off-the-shelf products will likely suffice. For processes that define your business – taking a specific approach to customer service, a unique troubleshooting methodology, or a particular way of handling escalations – custom agents that embody those processes can justify the investment.
Here are some tips:
Start small and focused. Don’t try to build a perfect AI assistant that handles everything; pick one narrow, high-impact use-case, as outlined earlier. If your competitive advantage is your specific approach to customer service, it’s worth investing in custom agent development; if it is a commodity process, look for an existing solution.
Document your processes. AI agents are only as good as the knowledge and processes they are given. If you cannot clearly define how something should be handled, an AI agent will not be able to do it consistently. Additionally, this process documentation exercise often reveals inefficiencies worth fixing regardless of AI.
Frame it correctly with your team. Remind them that this isn’t about replacing people; it is about removing repetitive work so they can focus on what humans do best. Morron emphasizes that the goal of AI agents is to free employees’ time to build relationships, solve problems, and perform the kind of complex thinking that builds businesses and customer loyalty.
Don’t wait. AI technology is rapidly improving and is already useful for narrow applications today. Waiting for it to be perfect means watching competitors gain experience and a competitive advantage while you sit on the sidelines.
About the Author
Paul Rothman
Editor-in-Chief/Security Business
Paul Rothman is Editor-in-Chief of Security Business magazine. Email him your comments and questions at [email protected]. Access the current issue, full archives and apply for a free subscription at www.securitybusinessmag.com.