Artificial intelligence can turn container images into inspection records, while connected sensors follow cargo beyond the port gate. But a port’s digital upgrade also depends on knowing who is inside the terminal, where people and equipment are moving and whether an emerging risk requires action.
A vessel enters the port carrying containers. The movements of trucks, equipment and workers are carefully coordinated. Then two vehicles collide—or an unauthorised person enters a restricted area—and the operating sequence begins to unravel.
For Pablo Aguirre, CEO of Stowlog, the problem is not necessarily an absence of procedures. Ports already have access rules, safety protocols and international security obligations. The gap often lies in their ability to see what is happening early enough to act.
“The question is not, do we have procedures? I think that the correct question is, do we have control?”
During the 29th Annual General Meeting of the Port Management Association of the Caribbean in Paramaribo, two successive sessions on smart-port technology and operational security examined this challenge from complementary perspectives. Leif Ollivierre, CEO of Naxxar Technology, focused on artificial intelligence and connected containers. Aguirre looked beyond the cargo to the people, equipment and movements inside the terminal.
Together, the sessions suggested that the next stage of port digitalisation will not simply generate more data. It will be judged by whether that data gives Caribbean ports clearer operational visibility.
AI gives the container a digital inspection record
Ollivierre’s presentation began with a constraint familiar to many Caribbean ports: their land, personnel, inspection capacity and terminal infrastructure cannot necessarily expand at the same pace as cargo demand.
Rather than reproducing the expensive inspection installations used by large international hubs, he presented an approach based on equipment that ports may already possess—a mobile phone or cameras installed at terminal gates.
With the ScanBox AI system demonstrated during the session, an employee records the different sides of a container. Artificial intelligence processes the images and identifies possible dents, holes, rust or other visible changes. The platform can also assemble a digital representation that allows the user to examine different surfaces of the container.
Ollivierre said that images of a 40-foot container could be collected with a phone in between 30 seconds and one minute. This was a performance claimed during the demonstration, not an independently verified result across Caribbean terminals.
The more significant function is the ability to compare inspections conducted at different stages of an operation. A container could be recorded when it enters a terminal and inspected again when it leaves, returns or passes to another operator.
A difference between the two records may help establish when damage occurred and support discussions between the terminal, shipping line, haulier, container owner and insurer.
That digital comparison would not automatically constitute legal proof. Lighting, rain, dirt, image quality and camera position could influence what the system detects. A port would still need a standard inspection protocol, secure files and a process for human validation when the technology identifies an anomaly.
Connected sensors extend visibility beyond the port gate
The second technology layer follows the container after the visual inspection.
Ollivierre presented a connected-container platform capable of collecting or reporting information such as location, temperature, movement, impact and possible opening of the container. Photographs of the cargo, seal and loading process could be attached to the same digital record, together with commercial documents and inspection reports.
A regional shipment between Saint Lucia and Dominica was used to illustrate the model.
At origin, the container and its contents would be photographed during loading. The images could be time-stamped and geolocated, while the seal would also be documented. During the journey, the device could record movement and generate an alert if the container appeared to have been opened. A further inspection at destination would create a second record for comparison.
The practical value is not limited to watching a marker move across a map. It comes from linking several forms of information:
- where the container travelled;
- what was reportedly loaded;
- what condition it was in;
- whether an unusual event occurred;
- and what was observed at destination.
The Saint Lucia–Dominica example should nevertheless be treated as a scenario presented by the technology provider. The available material does not establish whether it represents a full commercial deployment, a pilot or a demonstration.
Connected devices also have operational limits. Network coverage can be interrupted, batteries can fail and an alert does not explain an event by itself. A sensor may show that a container was opened, but it cannot always establish who opened it, whether the intervention was authorised or whether any cargo was removed.
The device creates a signal. People and institutions must still interpret it.
More information could make inspections more selective
The presentation also connected container visibility with risk-based inspection.
Before the cargo arrives, authorised users could potentially consult its packing list, commercial invoice, loading images, seal information and the alerts generated during transport. A regular shipper with a consistent history could then be assessed differently from a consignment displaying an unexpected opening, an unusual alert or a discrepancy between the documents and the images.
This does not mean that artificial intelligence would replace Customs officers or release cargo automatically.
Its more realistic function would be to help authorities decide where limited inspection resources should be concentrated. Customs could prepare for a particular examination before arrival, identify inconsistencies earlier and give greater attention to higher-risk consignments.
The final decision would remain with the competent authority.
That distinction matters because the technology presentation linked digital information with the possibility of advance cargo processing. A supplier can demonstrate that documents, photographs and sensor data are available. Only governments and border agencies can determine whether those records satisfy the legal requirements for clearance.
A smart port must also know who is inside
Containers are only one part of the port environment.
Aguirre’s presentation turned the discussion towards employees, contractors, truck drivers, visitors, customers and other external personnel moving through the terminal.
Compliance with the International Ship and Port Facility Security Code already requires ports to control access to their facilities. Aguirre’s concern was whether those procedures give decision-makers a centralised, current view of who is present and why.
A person may have permission to enter a port without being authorised to access every part of the terminal. Effective control therefore requires more than confirming an identity at the gate. It may also involve knowing:
- the purpose of the visit;
- the area the person is allowed to enter;
- the expected route;
- and whether the individual has received the required safety information.
Aguirre argued that a truck driver, contractor or visitor unfamiliar with the terminal environment may behave unpredictably because they do not know its routes and risks. That behaviour can create an accident even when there is no malicious intent.
The digital upgrade, in this case, consists of connecting access authorisation with what happens after the person enters the facility.
Real-time movement can reveal emerging risks
Aguirre described a project involving the Port of New York and New Jersey in which a virtual representation of the facility brings people and port equipment into a shared operational environment.
According to the presentation, authorised users can see movements in real time and act when a person appears in an area where they should not be. Aguirre said the approach was being used across three facilities to support more standardised procedures.
These were results described by Stowlog, rather than an independent evaluation of the system’s performance. Even so, the use case raises an important question for Caribbean terminals: could better visibility help port teams identify an emerging operational risk before it becomes an interruption?
A contractor moving away from an authorised route could simply be lost. The same movement could also place that person near heavy equipment, interrupt an operation or require a security check.
The purpose of an alert should therefore be to prompt a proportionate response, not to treat every deviation as proof of a threat.
The same principle applies to the digitalisation of training and registration. Aguirre cited a South Florida example in which driver induction and access processes conducted at the terminal entrance contributed to queues. Moving some registration, verification and safety training online before arrival could reduce the work required at the gate.
This is where safety and efficiency can support each other. A driver who understands the terminal rules before arriving may complete the access process more quickly and behave more predictably inside the facility.
But the outcome depends on whether the training is understood, verified and connected to the physical access-control system.
The real test begins after the demonstration
A smartphone inspection, a connected sensor or a virtual terminal can appear relatively simple during a presentation. Implementation is more complicated.
The first test is integration. The tools may need to exchange information with the Terminal Operating System, Port Community System, Maritime Single Window, Customs platform and existing access-control or security systems. A new platform that cannot communicate with them may simply create another isolated source of data.
The second is governance. Ports must determine who owns the photographs and tracking records, where they are stored, who may modify them and how long they should be retained. Cargo data can be commercially sensitive, while the continuous tracking of employees and contractors raises questions about privacy and proportionality.
The third is total cost. A solution based on phones and small sensors may require less physical infrastructure than a large scanning gate, but the equipment price is only one element. Connectivity, software subscriptions, integration, cybersecurity, maintenance, replacement devices and employee training must also be included.
Responsibility must remain clear. A port needs to know what happens if the AI misses visible damage, a sensor stops transmitting or a system fails to flag an unusual movement.
Technology should support accountability rather than disperse it among the port, operator, user and supplier.
Finally, adoption requires organisational change. Aguirre identified resistance to change as one of the central barriers to digitalisation. Employees need to understand why information is being collected, how it will be used and where human judgement remains essential.
A port does not become smart simply because its containers, workers and equipment generate more data. The upgrade becomes meaningful when that information is reliable, accessible to the right people and connected to clearly defined operational decisions.
