Reflections: What the NATO Ankara Summit means for Command-and-Control

Written by Andrew Graham, Group Senior Vice President Business Development

A fortnight on from the NATO Summit in Ankara, and I have been reflecting on what stood out most for me. The Summit Declaration sets the right direction: an interoperable warfighting cloud, powerful AI models, a stronger Europe within a stronger NATO. But every one of those ambitions rest on a single, less headline-grabbing capability, Allies need to be able to operate together - interoperability. And it is the one commitment in the declaration that is not a promise but a proven fact.

Most of the declaration is written in the future tense, and rightly so, because these are hard things still to build. One line is different: "Our investments are delivering the capabilities we need." Delivery is the real test, and in C2 it has largely been met, not in a demonstration but under the conditions that decide a coalition operation: many nations working together, degraded communications, command exercised by an officer from another nation and in using systems that were not initially designed to work together.

Interoperability cannot be asserted, only demonstrated, and across the Alliance it already is – in particular in the land domain. NATO chose SitaWare Headquarters as its Land C2 capability two years ago, and now more than 20 NATO members also use the same software as their national solution. This ensures that they are not only all trained on the same software, but they can instinctively work as one cohesive unit. Another example is the 1 German-Netherlands Corps which commands 12 nations from one operational picture, each keeping its own systems and rules while feeding it. This is tested where it counts: every year at CWIX against Alliance data standards, among them MIP 4.4, APP-11 and Federated Mission Networking (FMN). And more tellingly in coalition exercises, such as Aurora 26 earlier in the spring. Coalition interoperability at this scale is the hardest thing in defence software to fake – it is either there or it isn’t. And when it isn’t, this makes working as Allies extremely difficult.

The declaration puts weight on resilience too, and in command and control that is above all an operational quality. Capability has to keep working when networks are degraded or contested, not only when conditions are ideal. And it has to keep meeting the Alliance's standards continuously, rather than being certified once and left to drift. Interoperability that endures under those conditions is the point.

The same logic decides how NATO should deploy AI. Battlefield data already outruns what any staff can process unaided, so the question is not whether to use AI but where it sits. It belongs inside the C4ISR environment, drawing on the C2 and interoperability the Alliance already runs on, not bolted alongside as a parallel system. Used this way, AI helps staff find what matters, while the decision and the accountability remain with the commander, who must always be able to trace and question it. Agentic AI, the step beyond, must be built on those proven foundations, not in their place.

In C2, the interoperability declaration is fundamental, it is how the Alliance operates today and the foundation Agentic AI must build on. As a NATO partner, we are committed to continually develop and strengthen this interoperability foundation.