On February 4 and 5, 2026, Proportione presented Avian —our real-time bird collision risk prediction system— in the III Summit on Responsible AI in the Military (REAIM 2026), held at the PALEXCO Conference Centre in A Coruña. It was the first time that a Spanish technology consultancy of our size presented an AI operating system applied to aviation safety to representatives of more than 50 countries.
A "Waze of the air" to prevent collisions with birds
Collisions between birds and aircraft (bird strikes) are one of the most underestimated risk factors in aviation. According to the FAA database, more than 17,000 incidents are recorded annually in the United States alone, at an estimated cost of $1.2 billion per year. In Spain, the problem is especially acute due to the geographical position of the Iberian Peninsula as a migratory corridor between Europe and Africa.
Proportione Aviaria addresses this problem with an approach we've dubbed internally "the Waze of the air": a system that predicts where there will be concentrations of birds before a plane takes off, allowing you to plan safer routes or adjust flight schedules. It is not a reactive system that detects birds on radar; It's a system Predictive that anticipates risk.
How it works: Heat maps, AI, and a co-pilot who doesn't decide for you
The system is based on three technical pillars:
- Dynamic heat maps on hexagonal grid H3: we divided the airspace into hexagonal cells (the same standard Uber uses to optimize collections) and calculated the probability of bird presence in each cell, hour by hour, incorporating meteorological, seasonal and ornithological data.
- AI Prediction Engine: Machine learning models trained with historical data of sightings, migratory routes, atmospheric conditions and records of previous incidents. The system learns and recalibrates with each new piece of data.
- AI Co-Pilot with 20-60-20 Protocol: Artificial intelligence does not make decisions. It assists the controller or pilot by presenting risk information in a clear and actionable way: 20% of the critical context, 60% of the analysis generated by AI, and the final 20% of criteria that always remain in human hands. It is our responsible AI model, the same one we apply in education and consulting.
The entire system operates on infrastructure On-premise — that is, within the customer's premises, without sending sensitive data to the cloud — which is a non-negotiable requirement in the defense space.
What the pilots said
During the two days of the summit, our stand was visited by dozens of professionals from different countries: military and civilian pilots, air traffic controllers, flight safety managers and representatives of defense industries. The feedback was direct and consistent:
- Several military pilots noted that Mission planning today does not incorporate real-time ornithological data systematically. What exists are static NOTAMs and personal experience. Aviaria seemed to them a qualitative leap.
- The controllers particularly valued the Protocol 20-60-20: that the AI does not try to substitute its judgment, but presents the information in a way that facilitates the decision. "It's exactly what we need: data, not orders," a controller from a NATO country told us.
- Civil aviation representatives showed interest in Adapt the system to commercial airports, where the bird strikes in the approach and take-off phases they concentrate the greatest risk.
The Minister of Defence visited our stand
One of the highlights of the summit was the visit of the Minister of Defence, Margarita Robles, to our exhibition area. Mayte Tortosa, Aviaria's project manager, had the opportunity to present to her in person how the system works: heat maps, prediction logic and the responsible AI approach that underpins it.
The Minister was particularly interested in two aspects: how to ensure that data remains within the customer's infrastructure, and how the 20-60-20 protocol is integrated with existing in-flight risk management procedures. Both are, precisely, the two points that differentiate us the most: Data security and human augmentation, not blind automation.
Why a consulting firm like Proportione is in REAIM
REAIM is not a typical technology fair. It is the largest international forum on Responsible use of artificial intelligence in the military, promoted by the Government of the Netherlands and co-organised with Spain in this third edition. A Coruña brought together ministers, chiefs of staff, researchers and technology companies from more than 50 countries to discuss how AI should be used – and how not – in defence contexts.
Proportione was there because Aviaria is a concrete example of what we stand for: Technology that amplifies human judgment rather than replacing it. We do not develop autonomous weapons. We develop tools that help people make better decisions in situations where information is complex, time is limited, and error has consequences.
That principle—which we call the Model 20-60-20— is the same one we apply when we help a company use Google search data to anticipate demand, or when we design an AI-assisted critical thinking system for a university, or when we build the complete digital presence of Fundación porqueViven. The context changes; the approach, no.
What comes next?
REAIM has been a turning point for Aviaria. The conversations initiated in A Coruña are opening doors to validation pilots in new scenarios. The system, which is already operating in production, is ready to scale. And the interest shown by both the military and civil aviation sectors confirms that the problem is real, the solution is necessary and the approach – responsible AI, On-premise, with the human always at the center—is the right one.
If you work in aviation safety, aviation risk management, or defense innovation and want to explore what Aviaria can do in your context, Write to us. We are actively looking for validation partners.
