SenTRACK 2.0 launches to deliver adaptive AI surveillance

SenSen has introduced SenTRACK 2.0, a generative AI-powered platform designed to enhance surveillance by providing adaptive, role-based automation and situational intelligence.

Traditional surveillance systems have long relied on fixed rules and predefined triggers, such as setting off alarms when a person enters a restricted zone or when a vehicle overstays its allotted time. While effective in controlled circumstances, these static, rule-based approaches face challenges in the context of increasingly complex public spaces and operational environments.

SenTRACK 2.0 is intended to address these challenges by interpreting behaviour in context and responding to real-time developments. The system functions beyond the limitations of conventional rule-based surveillance, which "rely on clear instructions and predictable inputs. You define a condition, something like: 'if movement occurs after hours in this area,' and pair it with a consequence, like a triggered alert or lockdown. The simplicity of that model is its appeal. It's easy to implement, easy to explain. But it comes with serious trade-offs. Rules, by nature, do not adapt. They don't interpret context. A child running might trigger the same alert as a genuine threat. An accidental fall may be flagged as trespass. A sneeze in a kitchen might go completely unnoticed, until someone gets sick."

SenSen describes these limitations as particularly problematic in critical sectors such as food processing, logistics and public transport. According to the release, "In mission-critical environments like food processing, logistics hubs, or public transit, those are not minor errors. They're operational and reputational risks. It's also resource-heavy. As your use of the systems grows, you need more and more people to sift through the tsunami of alerts. Finally, and arguably most importantly: rules-based systems don't learn. In high-stakes environments, such as transport hubs, central kitchens, civic spaces, that's a fatal flaw. They require constant manual rule-tuning and never truly improve."

SenTRACK 2.0 is built around role-based automation enabled by adaptive AI agents that "operate not as passive detectors, but as active participants, with roles such as a trained patrol officer, quality inspector, or safety manager. It watches with purpose, asking not just 'did someone cross a line?' but 'what are they doing, and is it acceptable for this context and time?'"

The platform uses vision-language models (VLMs) and generative AI to "understand behaviour. It recognises patterns like aggression, contamination risk, collapse, or distress—even when they don't map neatly to a rule. It responds in real time, escalates intelligently, and offers natural-language interfaces that allow operators to query footage or create custom alerts using plain English."

SenSen claims that the system "behaves like a human inspector. That means that it's always present, always focused, and always learning." This ongoing adaptation includes the ability to adjust as environments change, such as when a new walkway opens or hygiene measures are updated. "It begins to understand what 'normal' looks like for a specific role, task, or location, and flags anything that deviates. No waiting. No retraining. No blind spots."

This shift is positioned as moving surveillance from reactive to proactive. "Teams aren't just reacting to events but rather staying ahead of them. In environments where risk is ever-present and standards are non-negotiable, such as in food safety, logistics, or civic operations, this shift from reactive to proactive can be the difference between compliance and catastrophe. SenTRACK 2.0 doesn't just observe. It provides situational awareness. The same system that flags a security breach can also surface operational inefficiencies. Where bottlenecks form. Where tasks are skipped. When fatigue is setting in. Why incidents keep occurring in the same zone."

The release emphasises the importance of such layered insights: "This kind of layered insight is critical in multi-role environments, from central kitchens to train platforms. A system that learns across these domains doesn't just prevent failure. It also helps organisations operate better."

The company states that integration is streamlined, noting that "SenTRACK integrates into your existing environment. It runs on the edge. It continues working even if connectivity drops. It doesn't require infrastructure overhauls. And it delivers insight in seconds and not hours or days."

SenSen acknowledges that while role-based automation is not flawless and requires thoughtful deployment, it offers significant benefits. "Is role-based automation perfect? No system is. It demands smarter AI models, robust hardware, and thoughtful deployment. But the upside is significant. Its limitations aren't in what it can't do, but rather in how quickly it continues to improve. This is a trade-up, not a trade-off. SenTRACK 2.0's edge processing, natural-language UI, and role-based architecture give control rooms and compliance teams something they've never had before: a system that reasons. That sees intention. That acts with context."

The release concludes by highlighting the shift from rule-based to role-based models: "As our urban, industrial, and regulatory environments become more complex, the cost of missing what matters is rising. Rule-based models were built for static, predictable environments. Role-based intelligence is built for the real world. One filled with people, messiness, urgency, and meaning. Whether it's ensuring safety in a public square, hygiene in a prep zone, or compliance on a cold chain dock, SenTRACK brings clarity where others bring footage. Because in the real world, intelligence makes all the difference."