From data to the simulator: 40º West’s case for smarter airline training

Airlines are collecting more information about how their crews perform than ever before. The challenge, according to 40º West CEO Borja Diaz Capelli, is turning that information into better training.

For Diaz Capelli, that means connecting what happens in the aircraft, classroom and simulator rather than treating training records, flight data and safety information as separate datasets.

“The data helps us with the way that we have to train. It’s not the rule maker,” he says.

Today, 40º West is focused on developing training management systems around the way airlines actually operate and train. Rather than starting with standard software and asking an airline to adapt its processes around it, the company examines how an operator actually trains, before configuring and tailoring the platform around its requirements.

After almost four years of development, that approach is now being deployed with airline customers, while the platform continues to evolve through ongoing R&D and feedback from users. The emphasis is on continuous iteration: developing the technology alongside the airlines using it, rather than treating the TMS as a finished, off-the-shelf product.

The TMS extends beyond data analysis and evidence-based training (EBT). It brings together functions including training coordination, assessments and competency tracking, lesson planning, instructor workflows, simulator profiles, documentation and reporting, while integrating with an airline’s existing systems. The aim is to connect the different elements of training within a single environment rather than managing them through separate systems and datasets. Typically, 40º West spends around a month examining an operator’s processes before configuration and implementation begins.

Putting data into training

That philosophy is currently being deployed at scale with Air Europa, where 40º West has spent the past year implementing a new training management system.

The project has involved integrating the platform with existing systems, one of which dates from the 1990s. Diaz Capelli says the implementation has reduced manual processes, improved data integrity and replaced large quantities of paperwork with digital records.

More importantly, it is allowing the airline to identify where training needs to change.  40º West’s platform can integrate information from areas including flight data monitoring (FDM), fatigue risk management, safety, maintenance and training. Operators can use it to identify operational trends that could become training priorities, and subsequently examine whether performance changes after the training intervention. An unstable-approach trend, for example, could inform a simulator scenario, with subsequent data showing whether performance improves.

Previous performance can also influence what a pilot encounters during training. If a pilot has struggled with a particular approach during previous simulator sessions, it can be included again. If theoretical assessments identify a weaker knowledge area, additional questions can target that subject.

Air Europa has also used operational data to examine how 737 crews apply brakes, with braking technique incorporated into simulator training. Data relating to 787 climb performance has similarly been fed into training.

The objective is not simply to perform a manoeuvre safely, Diaz Capelli says, but to consider how effectively it is performed, potentially connecting training with wider operational priorities such as fuel efficiency.

It reflects a broader change in airline training. “In the past, we were trained like a bot,” he says, describing recurrent programmes built around familiar sequences of engine failures, fires and other prescribed exercises. Today crews can encounter scenarios derived from events experienced by their own airline or elsewhere in the industry, allowing them to examine how real crews responded and why.

Beyond ticking the CBTA box

The shift is closely aligned with Evidence-Based Training (EBT) and Competency-Based Training and Assessment (CBTA). But Diaz Capelli cautions against assuming that adopting CBTA is simply a matter of changing an airline’s training programme.

One of the biggest challenges is instructor development. An instructor new to CBTA might observe that a procedure has not been completed correctly and mark down the relevant procedural competency. A more experienced instructor may look further and discover that the underlying problem was workload management rather than procedural knowledge.

An instructor assessing within a competency-based system for the first time will not necessarily interpret performance in the same way as one who has spent five years working with it. That becomes particularly important as airlines seek to personalise training while maintaining consistent assessment standards. For Diaz Capelli, instructor standardisation is central to reconciling the two.

What should AI do?

Technology can help, particularly with one of the richest but least structured sources of training information: instructor comments. Assessments generate scores for competencies, manoeuvres and malfunctions, but debriefing notes can contain observations that do not fit neatly into those categories. 40º West is using artificial intelligence to analyse those comments and turn them into information and KPIs. Instead of requiring a training manager to read large numbers of individual reports, AI can help identify recurring themes.

AI can also support adaptive knowledge testing. If an assessment indicates weakness in particular subjects, the system can dynamically introduce additional questions. But Diaz Capelli draws a clear line between using AI to process information and allowing it to make training judgements. AI can identify patterns and suggest where additional training might be needed, but the instructor still needs to decide what those patterns mean.

“We need to know the context,” Diaz Capelli says. “And it is very difficult for the AI to know the context.”

There is also a risk in giving instructors too much information about an individual before a session. Previous records could influence their expectations and potentially their assessment. The system can therefore use a much wider dataset to create a training profile without necessarily exposing all the underlying information to the instructor.

From technology to training

As aviation training becomes more data-driven, more information does not automatically mean better training. Operators still need to decide which information matters, how it should change what happens in the simulator and whether an intervention has worked. Diaz Capelli believes aviation expertise must remain at the centre of that process.

“When I sit with a training manager, I can talk with them face to face,” he says. “I’m not talking about a product. I’m talking about a reality.”

At EATS 2026, he plans to use the Air Europa implementation to demonstrate that argument, examining how the project has changed processes, training and operational efficiency. For Diaz Capelli, the future is not about replacing instructors with algorithms. It is about giving them better evidence on which to base their decisions.

To learn more, stop by Booth 211 during EATS, taking place this 3-5 November at Estoril Congress Centre, Cascais, Portugal.