Why Field Testing Matters in Defence AI

How live exercises, operator feedback, and real-world interoperability challenges shape TerraSense software development.

At TerraSense Analytics, regular participation in live exercises and field environments is central to our software design and refinement. These events serve as engineering environments where we test assumptions, stress architectures, and evaluate interoperability in real time.

Learning from real operational environments

Through multiple exercises and deployments, including ISC testing in Gagetown, NASP events in Gaspé, and live experimentation at EDGE 23 and PC-C4, we have found that operational relevance is only achieved through direct exposure to end-user workflows.

Each environment introduced distinct sensor systems and operational constraints, yet more importantly, these engagements exposed various Concepts of Operations (CONOPs) that revealed how soldiers and commanders actually employ systems under stress. From CFB Petawawa to CFB Edmonton, the variation in terrain and mission profiles has reinforced a critical lesson: interoperability is as much a human challenge as it is a technical one.

Interoperability is proven, not assumed

A recurring outcome of these engagements is the validation, or often the invalidation, of interoperability assumptions. While integrating new sensors appears straightforward on paper, the reality of the field quickly surfaces discrepancies in timing, bandwidth, and data labelling. What looks “compatible” in a specification often requires significant adaptation when exposed to real-world systems-of-systems environments. Adopting widely accepted, standard data formats is the critical preparatory step that minimizes these on-site surprises and accelerates deployment.

By working directly with allied forces and defence scientists, TerraSense ingests data from new sensor platforms under realistic conditions. This proves interoperability is enabled by a foundational commitment to data standardization, rather than last-minute patches in the field.

Understanding CONOPS through observation, not assumption

Perhaps the most valuable insight gained from these environments is not technical; it is behavioural.

Watching how sensor operators actually work under field conditions reveals gaps between intended and actual use, often highlighting where manual workarounds emerge or how time constraints shift priorities. Data interpretation changes based on the mission phase, and these nuances are easily missed in a controlled setting.

These observations directly inform our software design. We continuously refine our interfaces, data pipelines, and analytics to reflect how systems are used, not just how they were envisioned in design documents.

This is where engagement with organizations such as DRDC and allied experimentation communities becomes essential. Their willingness to expose systems to realistic testing environments enables a feedback loop that cannot be replicated in controlled lab-type settings.

Building credibility through repeated field exposure

Across EDGE 23, PC-C4, and multiple Canadian Armed Forces environments, a clear pattern has emerged: consistent participation in operational experimentation builds more than technical validation, it builds trust.

When end users see their input reflected in successive software iterations, and operators recognize their specific workflows in the interface, the technology begins to feel like a tool rather than a burden. In defence technology, credibility is not built through claims. It is built through a repeated presence in environments where failure is visible and improvement is measurable.

Why being on the ground matters

For TerraSense, being on the ground with end users ensures three things.

• First, systems are shaped by real mission requirements rather than assumptions.

• Second, edge cases and operational stressors are identified early before they become critical failures.

• Finally, integration is tested across real platforms instead of theoretical interfaces.

Every event, every exercise, and every deployment contributes to a clearer understanding of how to build systems that function where it matters most: in the field.

Closing thoughts

Defence innovation does not happen in isolation. It happens in collaboration with operators, scientists, and allied partners who are willing to test systems in real environments and provide unfiltered feedback.

Organizations such as the Canadian Army Land Warfare Centre, DRDC, DLCSPM, CANSOF and allied experimentation programs like EDGE and Project Convergence have played a critical role in enabling that process. Their continued commitment to experimentation ensures that emerging technologies are not just developed but are operationalized for those on the ground.