Cargo theft remains a persistent threat to the logistics industry, with incidents becoming more sophisticated and harder to stop. Milan Luketic, Chief Technology Officer at Birdseye Security Solutions, believes AI-powered surveillance can help flip the script—detecting threats early, streamlining gate operations, and potentially stopping crimes before they happen. In this Q&A, Luketic explains how AI is reshaping facility security and why it might be more cost-effective than hiring human guards.

What’s Related

Supply Chain 24/7: Cargo theft is a growing concern in transportation and logistics. From your vantage point, why is this still such a persistent issue?

Milan Luketic: Cargo theft has been rising in recent years, and the risk is especially high now as inventory levels fluctuate amid market volatility and tariff uncertainty. Many businesses hold more inventory than usual as they try to get ahead of tariffs, attracting more attention from bad actors. On top of that, theft generally surges during periods of economic downturn. 

We’ve also seen a rise in organized group attacks, where multiple individuals storm a logistics facility, making it harder to stop the theft. Criminals are becoming more sophisticated, using fake paperwork and exploiting weak gate security to gain access to facilities. 

But it’s hard to fathom the true scale of cargo theft, as many incidents go unreported due to reputational risk and concerns of insurance rate hikes.

SC247: Can you describe how AI can help prevent cargo theft?

ML: AI-powered systems are designed to enhance security in truck yards and facilities by providing automated surveillance and access control. The system uses an array of cameras along the perimeter and gates to detect intruders and monitor activity. The AI technology verifies truck numbers and authorized personnel, allowing for automated access or requiring human verification when necessary. This reduces the need for physical security guards and increases efficiency in managing entries and exits.

AI helps prevent cargo theft by continuously monitoring the yard with multiple cameras, detecting unauthorized attempts to enter, and raising alarms when suspicious activity is observed. The system also integrates with yard management platforms to pre-approve entries, making the process faster and more secure.

SC247: How do AI-powered solutions complement or replace traditional on-site guards or surveillance staff?

ML: AI-powered solutions like Birdseye can complement or replace traditional on-site guards or surveillance staff by providing more efficient and comprehensive security measures. In terms of intruder detection, having 120 AI cameras stationed along the fence line replaces the need for a guard to monitor several screens simultaneously. The AI can perform behavioral analysis, detecting suspicious behavior such as climbing a fence, and then escalate using deterring measures. 

Regarding gate operations, AI can automatically verify truck numbers or driver identification, allowing authorized vehicles to enter or exit the yard without manual checks by guards. This touchless transaction speeds up the process at access points and reduces human error. That said, human agents must still be involved to handle complex cases or tasks requiring judgment, including facial recognition or incidents of fraud, such as impersonation attempts. 

SC247: You mention the shift from reactive to proactive. Can you explain what that actually looks like in practice?

ML: Traditionally, cargo theft attempts often are not detected until it’s too late to deter the crime. For example, a guard may be on the eighth hour of their shift, trying to watch 50 camera screens while also processing a growing line of frustrated truck drivers waiting at the gate. If an intruder were to jump the fence, the guard likely wouldn’t be monitoring that precise screen the second it happened, meaning the threat could go unnoticed until the intruder plows through the gate with a stolen truckload. At that point, the guard’s hands are tied – it’s difficult for them to stop the crime without risking their personal safety and liability. 

AI, on the other hand, would detect and flag the intruder the second they get too close to the fence. This enables a proactive, preventative response—a law enforcement presence can deter the crime before it escalates.

Another (less nefarious) example facility operators often confront is a driver mistakenly taking the wrong truckload. Instead of discovering the issue after the truckload fails to arrive at the right warehouse at the expected time, AI could cross-check the truck’s serial number before it exits the gate and flag if it doesn’t match what’s in the yard management system. In both cases, AI helps prevent losses of time and money. 

SC247: What alerts or automated actions can be triggered when a threat is detected?

ML: When a threat is detected, Birdseye’s system can trigger several types of alerts and automated actions to ensure an appropriate response based on the level of risk.

If a pedestrian is too close to the fence at the low end of the spectrum, a strobe light will go off as an initial deterrent. If the person remains near the fence, a loudspeaker will issue a voice warning, instructing them to step away or indicating that the police will be called. If the person persists, the system will escalate the situation by calling the police. 

SC247: How do you ensure your systems can differentiate between a false alarm and a real threat?

ML: We ensure our systems differentiate between false alarms and real threats through AI and human intervention. Our AI is trained to detect behavioral intent, such as someone climbing a fence, which would trigger an alarm. In situations where AI is unsure, human agents are involved in making the final judgment. 

SC247: How do your customers typically calculate ROI when investing in AI-based surveillance and security?

ML: Our customers typically calculate ROI by comparing the hourly cost of the security guards they have on-site to the hourly cost of a remote live video monitoring system (which includes the AI and our professionally trained remote agents). This direct cost comparison often yields significant savings. Many customers can also lower their insurance costs by demonstrating better security practices and proactively mitigating risks. Some customers also consider ROI metrics related to efficiency, such as improved driver retention due to faster gate transactions and reduced wait time. 

SC247: Have you seen any success stories or examples of an incident where your AI system stopped it before it escalated?

ML: On many occasions, AI has alerted our agents to smoldering on the security perimeter. This has resulted in an agent’s call to the fire department, who arrived with plenty of time to snuff out the growing flames before they become a full-fledged fire that threatens personal safety or causes property damage.     

SC247: What advice would you give to logistics or transportation companies just beginning to evaluate AI-based security solutions?

ML: Get hands-on. Engage a vendor and ask for a demo or on-site trial. You will learn so much by applying AI technology to a facility you’re already familiar with and discover what’s possible and the limits.   

SC247: If a company doesn’t have a large security budget, are there entry-level options to start with AI?

ML: Yes – in fact, AI can actually be more cost-effective than hiring human security guards for companies with a limited budget. A human guard typically costs about $25-40 per hour (depending on the state), and many facilities have two guards staffed on each shift. If the company were to add 100 AI cameras, it could substantially reduce its labor costs because it’d only need one human agent who could monitor multiple sites remotely by reacting to AI-generated alerts. This can amount to thousands of dollars in cost savings each month.

Milan Luketic is Chief Technology Officer at Birdseye Security Solutions.