The problem
Automated suppression systems are often deployed into environments that are operationally busy and physically challenging: commercial kitchens with heat and grease, warehouses with long runs and mixed materials, and manufacturing sites with noise, vibration, and moving equipment. The suppression hardware itself can be highly effective, but operators still face a persistent reality: if you cannot see the system state in real time, you end up relying on periodic checks, manual reporting, and delayed awareness of faults or abnormal conditions.
The firm’s customers wanted stronger confidence between inspections. They wanted to know whether a unit was healthy, whether a canister pressure was drifting, whether temperature behaviour was trending towards risk, and whether any abnormal conditions were present that required action. In practice, that translates into a simple operational request: reduce uncertainty. If something is going wrong, alert early enough that there is time to respond calmly rather than in a scramble after an incident.
There was also a commercial reality. The manufacturer was scaling deployments across multiple customer organisations and site types. A single-tenant dashboard or a one-off install script does not survive that transition. They needed a proper multi-tenant platform that could support resellers, service partners, and end customers with distinct access and clear separation, while still providing an engineering-grade view for diagnostics when issues arise.
Scoping and constraints
Squared Technologies scoped the work as a full-stack delivery. That included embedded firmware for sensor and telemetry nodes, low-power RF communications in the sub-GHz range, gateway behaviour, ingestion and processing, alerting logic, dashboards, and the operational model for onboarding and support. The scope was outcome-led: reliable monitoring, meaningful alerts, and a platform that could be operated by real teams without specialist intervention during routine use.
The constraints were practical. Connectivity in the field is not uniform. Some sites have strong network infrastructure, others do not. Some environments are RF-hostile, with dense structure and interference sources. Sensors behave differently in kitchens than in cold warehouses. And because this is safety-adjacent technology, failure modes must be treated with discipline. The system must degrade safely and transparently. It must report its own health. It must avoid “silent failure” where everything looks normal until it is not.
We also treated data quality as a first-class requirement. Temperature analytics are only valuable if the underlying signals are stable and the interpretation is defensible. Pressure monitoring is only meaningful if you can trust calibration, sampling, and the logic that decides what constitutes a drift worth notifying about. The platform had to produce operational truth, not noise.
Delivery approach
The delivery began with consulting-led discovery. We mapped how the suppression product is installed, maintained, and serviced, and how the manufacturer’s customers actually respond to alerts. This matters because alerting is never just a technical function. It is a workflow. The only alerts that matter are the ones that cause correct action, consistently, without escalating fatigue. That informed the thresholds, the notification semantics, and the way events are presented in dashboards.
At the edge, the monitoring nodes were designed for low-power operation with sub-GHz telemetry. The 433 MHz band was selected as a key part of the communications strategy because it provides practical propagation benefits in buildings and across mixed environments. The objective was not theoretical range figures, but real deployments where signals need to traverse walls, plant rooms, and busy operating spaces. Low power matters too: it reduces maintenance burden and supports a product that can be installed widely without becoming an operational headache.
On-device firmware handled sampling, filtering, and the creation of clean, timestamped events. Temperature signals, for example, can exhibit noise and transient spikes depending on placement and local conditions. Pressure signals can drift slowly over time or show behaviour changes under temperature variation. We designed the embedded layer to produce stable telemetry, with clear event semantics, rather than pushing raw noise into the cloud and hoping dashboards will compensate.
Pilot, range testing, and fine-tuning
The programme moved quickly into a pilot phase. The goal was to validate end-to-end behaviour in representative conditions: a mix of building types, different install geometries, and operationally realistic constraints. We used the pilot to test not only whether data arrived, but whether the data was useful and actionable. That includes latency, false positive resistance, health reporting, and how quickly the system helps a service team narrow down an issue when something abnormal is detected.
Range testing followed as a structured engineering exercise rather than a marketing exercise. We validated performance across typical building fabric and site layouts, tuned the telemetry cadence to balance timeliness with power budgets, and refined gateway placement guidance so installations could be repeated confidently. The result was a system that behaves consistently across the environments the manufacturer actually sells into, not just idealised test conditions.
Fine-tuning focused heavily on alert quality. For a system to be trusted, it must be selective. We refined the temperature analytics and pressure drift detection so that alerts correspond to meaningful changes rather than harmless variation. We also ensured that when alerts do fire, they are contextual. A notification should tell an operator what changed, where it changed, and what the recommended response pathway is, without requiring a specialist to interpret a cryptic signal.
The platform
The platform was delivered as a multi-tenant system designed for real commercial rollout. Tenancy boundaries were treated as core architecture, not an afterthought. The manufacturer could provision customers, customers could manage their sites, and service teams could access the information required to support maintenance and diagnostics, all without compromising separation between organisations.
Dashboards were built to prioritise operational clarity. Operators can see system state across estates, identify units that require attention, and review event history with timestamped context. The backend supports secure ingestion, storage, and retrieval of telemetry, with alerting workflows that can route notifications appropriately. This was full-stack delivery: front end, back end, and the integration layer that makes real-time monitoring feel natural rather than bolted on.
We also engineered for ongoing operations. Device health, last-seen telemetry, and integrity signals are essential when equipment is deployed into busy environments. A monitoring system that cannot detect its own blind spots creates a false sense of security. The platform therefore treats health reporting as a first-class element of system state. If a node is not reporting, that is a condition to be handled explicitly, not a quiet assumption.
Conclusion
The delivered outcome is a suppression monitoring stack that behaves like a modern real-time system: timely detection of abnormal conditions, reliable telemetry across challenging environments, and an operational view that supports customers and service teams at scale. The manufacturer moved from a product that was effective in the moment of activation to a product that is visible and manageable every day, across many sites.
We avoid absolute claims, because serious engineering is honest about constraints. What we can say is that the system was designed and delivered to be fast, accurate, and reliable in real deployments, and to scale commercially without losing operational integrity. That combination is what differentiates a demo from a product that survives the field.
If you are building safety-adjacent systems and want a partner that can deliver firmware, RF telemetry, secure ingestion, alerting, and a multi-tenant platform end to end, talk to engineering. We scope around outcomes, design for the environment you actually deploy into, and build the stack as one coherent system.