Digital Content Editor, Eve Goode spoke exclusively with Michael Matta, CEO and Co-Founder of Solink about the ongoing impact of deploying AI agents.
How are AI agents being used to identify anomalies and flag potential security threats in real time?
The biggest shift is that video AI has moved beyond being able to create patterns from pixels, but rather bring intelligence to the forefront.
Traditional Computer Vision was good at classifying pixels in images/video to identify objects or use-cases (like people in a pre-defined queue).
Today’s AI agents can understand business context, which means they can support teams by acting as digital teammates that can intelligently assess risk or revenue opportunities.
Solink AI Agents monitor hundreds of cameras and data points in real time building an intelligent pattern of “normal” – the solution will surface specific moments that require intervention.
No longer will teams need to watch every camera or manually review every incident or transaction.
Instead, the right insights will surface, and they’ll be able to move from a reactive and investigative approach, to proactive awareness.
What are the biggest challenges enterprises face when deploying autonomous AI agents in physical environments?
The biggest challenge isn’t technology, it’s changing behavior.
AI is most valuable when it enables practitioners to spend less time watching video and more time doing, or taking meaningful action..
Video alone rarely tells the complete story.
To determine whether an event represents normal activity, an operational issue, or a genuine security risk, an AI agent needs to reason across video and signals from systems such as point of sale, access control and alarms.
AI Agents should show what was detected, why it matters, and what informed the decision.
Autonomous should not mean unsupervised.
At Solink, our goal is to extend the reach of experienced professionals, not replace their judgment.
How is AI-driven video intelligence changing the way retail and enterprise security teams approach loss prevention?
AI is fundamentally changing loss prevention by making it possible to identify patterns, anomalies and emerging risks at a scale that simply wasn’t practical before.
Traditionally, teams had to define every scenario they wanted to detect and manually investigate incidents after losses occurred.
Today, AI can recognise when something isn’t normal, even if it hasn’t seen that exact scenario before, helping organisations uncover risks much earlier.
That’s important because loss isn’t linear.
Small issues compound over time.
The earlier you identify and address them, the greater the impact on protecting margins and improving business performance.
What role can AI agents play in helping security professionals detect and respond to incidents more quickly?
AI agents fundamentally change the role of security teams by shifting their focus from investigation to action.
Instead of asking people to manually review transactions, footage or alerts, AI continuously observes what no human team could ever monitor at scale, learning what’s normal and identifying emerging patterns that signal risk.
Over time, those patterns become increasingly valuable, helping organisations detect incidents earlier, respond faster, and uncover risks they may never have known existed.
The result is that security professionals spend less time watching and more time preventing loss.
How can organisations evaluate whether AI-powered security solutions are ready to be deployed at scale?
The best way to evaluate AI is to start with a real business problem, not a blank slate.
Today’s AI is already mature enough to automate time-consuming work like investigations and video review, delivering measurable ROI without requiring organisations to replace their existing technology or create a new budget.
As AI agents continue to evolve, they’ll take on even more context and decision-making, but organisations don’t need to wait for that future.
Start by solving a meaningful problem today, measure the outcome and hold AI to the same or an even higher standard as any other business investment.
Looking ahead, how do you see AI agents transforming the future of enterprise security, and what should organisations be doing now to prepare?
The future of enterprise security isn’t about adding more people to watch more video.
It’s about pairing every security professional with AI agents that take on the repetitive work they do every day.
Organisations should start by mapping their current security team and identifying the manual tasks that can be handled by an AI agent, whether that’s monitoring cameras, reviewing transactions, or investigating routine incidents.
That allows security teams to move from being watchers to doers, focusing on prevention, response and higher-value work.
Just as importantly, evaluate AI against your total operating budget, not just your technology budget.
The real value comes from increasing the impact of every employee, not simply deploying another tool.