In June 2026, EASA published Decision ED 2026/004/R. It implements Regulation (EU) 2025/2143, which amends the rules every European air navigation service provider operates under. The headline is a date. From January 2029, controller training and assessment has to be competency-based, and the way a controller is judged competent changes with it.
At the centre of the new framework sits the adapted competency model, the ACM. It names the competencies a controller has to demonstrate, the observable behaviours behind each one, the conditions an assessment exercise must meet, and the standard that decides pass or fail. For an aerodrome rating, that’s ten competencies, from situational awareness to separation and conflict resolution, each with its own set of observable behaviours, assessed in an exercise that has to carry a defined traffic mix, a minimum number of movements, specific traffic scenarios, and at least one non-routine situation.
The standard is where it gets demanding. Safety comes first, so a single unsafe outcome caused by the candidate fails the exercise whatever else went well. The competencies don’t average out. A real gap on separation and conflict resolution can’t be bought back by a strong showing elsewhere, because separation has to be demonstrated every time the situation calls for it. One good run isn’t enough either. A competency is judged across several exercises and several different contexts before anyone signs it off.
Most units still run this on spreadsheets and slide decks. That holds up until an auditor asks you to show that every competency was observed, across enough varied situations, with a verdict that follows a rule rather than an instructor’s read of the day. Rebuilding that from a folder of files is slow, and it’s fragile. The 2029 date looks distant, but building an ACM, mapping a syllabus to competencies, and producing the traceability behind each assessment takes longer than it sounds.
EDS is built around this chain. A designer composes exercises from performance objectives and particular control situations, and EDS keeps each one linked to the competencies it exercises. When an instructor assesses, EDS computes the verdict from what was actually observed, under the non-compensatory rule, with safety and separation handled the way the standard requires. The instructor observes and records, the system applies the rule, and that’s what makes a verdict defensible when someone reviews it later. From there, EDS assembles the traceability that shows the exercises cover the ACM, down to the conformity of each exercise against the conditions the regulation sets.
One distinction is easy to miss, and it shapes the whole product. The fixed grid EASA publishes applies to rating training, the initial phase run by training organisations. Operational training at a unit works differently. There, the provider builds its own competency model on the ICAO framework, Doc 9868 and Doc 10056, shaped to its real environment. EDS handles both, because a tool that only knew the textbook grid would be no use to the units that do most of the training.
This already runs in production. At the French Navy’s Lann-Bihoué base, EDS has cut the time to build a training exercise by roughly two thirds. Major David Le Luyer, who uses it there, puts it at three times faster. That’s the hardest part of the chain to fake, the daily work of designing exercises that actually hold together.
The organisations that treat 2029 as a paperwork problem will be the ones scrambling in 2028. The ones that treat it as a design problem, competencies first and traceability built in from the start, will have far less to prove when the auditor arrives. That’s the work EDS is for.
