An auditor opens the crew list, points to five names at random, and asks the same thing for each one: the date on which that person actually went down an evacuation slide, put out a fire, and operated every type of exit, and the name of whoever witnessed it. Most of the time, the exercise did take place. What is missing is the piece of evidence that proves it — and it is almost always missing at the same points in the chain.
A Training Record Does Not Prove a Competency, It Proves a Fact
The confusion is old, and it is costly. The training file does not attest that a cabin crew member knows how to do something: it attests that a specific fact occurred, on a given day, under the eyes of a named person. Competency is observed and checked in line; traceability, on the other hand, is created at the moment of the exercise, or it is never created at all.
A file entry is only enforceable if it answers four questions without relying on anyone’s memory. Who: an unambiguous identity, not a first name and a fleet number. What: the exact exercise, on which type, with which equipment, on which type of exit. When: a day, not a month or a spring training course. Attested by whom: a name, a signature, and above all a qualification valid on that date. The fourth question is the one everyone forgets, and it is the one that brings down entire series of entries at once.
The stakes are not merely paperwork. Cabin crew qualification is a condition of the flight’s compliance. In front of an auditor, a qualification that cannot be demonstrated is equivalent to a qualification that does not exist: the auditor cannot record what they cannot read. The reasoning may seem harsh, but it is exactly symmetrical to the one the operator itself applies when it refuses to dispatch an aircraft whose workcard has not been signed off.
A training record is not about skills—it’s proof that someone did something specific on a certain day, with a witness. If the paperwork doesn’t show all four key details (who, what, when, and who signed it), it’s useless. Missing just one part can invalidate the whole record.
An auditor checks a cabin crew member’s file and asks for proof of a fire extinguisher drill. The file shows a “recurrent training completed” stamp for the year, but no specific date or instructor’s name. The operator can’t prove the crew member actually did the drill, so the qualification is invalid—even if they did it.
The Inventory of Records, Building Block by Building Block
A cabin crew qualification is not a single document; it is a stack of building blocks with different owners, different cycles, and different storage locations. It is this dispersion that produces the gaps, far more than anyone’s negligence.
| Item | Cycle | Evidence that holds up | Where it gets lost |
|---|---|---|---|
| Cabin Crew Attestation (CCA) | Issued once, subject to uninterrupted practice | Verified copy, issuing authority and reference recorded at hiring | File opened in the rush of a course, copy never filed |
| Conversion training to the operator and type | Before the first line assignment | Named end-of-course attestation and a record of the exercises performed | Course shared between operators, record kept by the training centre |
| Differences and variants | At every configuration deviation | Explicit mention of the variant and equipment concerned | « It’s the same aircraft »: not a single line written |
| Familiarisation flight | Before assuming duties | Date, flight number, mention of supernumerary crew status | Absorbed into an ordinary line roster |
| Recurrent training, theoretical part | Annual cycle | Dated, named attendance record, linked to the programme actually covered that day | Collective sheet detached from the content of the session |
| Heavy practical exercises | Multi-year cycle | One date per exercise and per person | A single date for an entire group and an entire course |
| Refresher after a break in flying | Triggered by the absence | Return date, content covered, instructor identified | Refresher done in line, never formalised |
| Checks | As per the approved programme | Result, examiner’s identity, validity of their qualification | Result entered without the identity of whoever pronounced it |
Two clocks run in parallel on this stack, and they never line up cleanly. Recurrent training comes around every year; the heaviest practical exercises — those requiring a mock-up, a fire pit, or an evacuation slide — come around on a longer cycle, on the order of several years depending on the approved programme. A management tool that tracks only a single deadline will therefore show a file as green even though a practical exercise has lapsed.
For recurrent training, approved programmes generally follow the usual convention found in the rest of the regulation: a renewal completed within the window preceding the deadline renews it from the original date, without shifting it. The rule exists to prevent an operator’s calendar from slowly drifting, year after year. It has a formidable side effect: a course taken too early credits nothing extra, while the long cycle remains anchored to its own reference point. A cabin crew member perfectly up to date on recurrent training can therefore have let a practical exercise lapse for several years without any alert being triggered.
The Date of a Practical Exercise: The Weakest Point
The elements that bring files down are always the same, and they are precisely the only ones that are individual by nature: actually extinguishing a fire with representative equipment, donning a smoke hood, descending an evacuation slide, putting on a life jacket and deploying a life raft, operating each type of normal and emergency exit, and using portable oxygen. Everything else — theory, briefing, demonstration — is proven by presence. These are proven by an action, and an action can only be proven by someone who witnessed it.
Yet the practical session is organised to be efficient, not to be traceable. Twelve people rotate through four workshops; two arrive late because a flight ran over, one leaves before the end to report for duty, the slide is unavailable that day and the workshop is postponed. By the end of the day, the session sheet carries twelve names, a date, and a list of workshops. It is wrong for three people, and no one will know it in six months.
The trap: a collective sheet proves presence at a session, never the completion of each exercise by each person. As soon as a workshop is postponed, split, or missed, the sheet claims more than what actually happened — and it is the operator who signed it. The remedy is not administrative: a box per exercise and per name, checked on the spot, at the mock-up, not reconstructed that evening at a desk.
The Six Breaks in the Chain Where the Evidence Evaporates
Evidence does not disappear at random. It disappears at interfaces — that is, every time it changes hands, medium, or tool. Six of these recur with enough regularity to deserve being addressed one by one.
Subcontracting. The operator cannot delegate the demonstration of its own personnel’s qualification. What comes back from an outside organisation is almost always an end-of-course attestation — a summary — while the detailed, exercise-by-exercise record stays in the provider’s own files. The day the operator changes provider, the granularity leaves with it. The safeguard is contractual: name the detailed record as a deliverable, with dates per exercise and the identity of the instructors.
The instructor’s paper form. Filled in at the mock-up, by hand, often standing up. Handed to a coordinator, scanned in batches. A batch fails, a box goes missing during a move, a sheet stays in a folder. This is not professional misconduct, it is handling — and handling without an acknowledgement of receipt always loses something.
Data entry. This is the most destructive break, because it is invisible and permanent. The management tool offers a single field, « recurrent completed, » which swallows eight distinct elements. The person entering the data does not lose information through inattention: they lose it because the form is not built to receive it. No after-the-fact check will bring it back. The only remedy lies upstream, in the structure of the field.
Tool migration. A change of crew management system is almost always done by migrating a status — valid until such-and-such a date — rather than the dated events that underpin it. The new tool is accurate for the present and silent about the past. It is precisely the past that the auditor asks about, which is why a migration must be accompanied by a readable freeze of the prior history, even in the form of a flat archive.
Transfer. Arrival from another operator, return from secondment, or crew provided under a wet lease arrangement: the prior file arrives summarised in a letter. A summary is not evidence. One must either obtain the records, redo the exercises, or state in black and white that one is starting from zero on the undocumented points — the third option is the least comfortable, and the only one that holds up.
Departure and archiving. The retention obligation outlives the employment contract, for a duration set by the record-keeping rules and repeated in the manual. A file archived on the day of departure often loses the final weeks — precisely the most recent period of activity, the one that will be questioned if an occurrence is traced back.
The Triangle That Makes an Entry Enforceable
A file line does not stand on its own. It holds because three records correspond to one another, and one can move between them in both directions: the individual file, the session sheet, and the qualification of whoever signed it. Remove one side, and the entry still exists but no longer demonstrates anything.
The side that no one checks is the bottom one. An instructor also has a validity, a standardisation, and a scope of qualification defined by the operator. If they signed after any one of the three had expired, the exercise may have been conducted perfectly and the attestation remains worthless: it was given by someone the operator no longer authorised to give it. This defect has an unpleasant property — it is never isolated. It falls in batches of files, all those bearing the same signature within the same time window.
The right reflex: pull a single thread, in both directions. Take a name at random from the system extraction and trace it back to the ink: the sheet, the session, the instructor, their qualification on that date. Then do the reverse — start from a session sheet and check that each name on it can be found, exercise by exercise, in the individual files. Allow half a day per fleet, twice a year. This test finds in a single morning what an internal checklist audit never sees, because a checklist verifies that procedures exist, not the continuity of facts.
When the Evidence Has Truly Disappeared
It happens that the exercise did take place and nothing proves it. The temptation is well known: pull out a sheet, have it signed by the instructor who remembers it perfectly well, and date it the day of the session. This is the one truly serious mistake in this whole subject.
Never sign retroactively. A signature affixed today under an old date regularises nothing: it falsifies a record. It turns a record-keeping discrepancy — correctable, with an action plan and a deadline — into an integrity problem that affects the organisation itself and personally implicates the signatory. These documents are easy to spot: uniform ink, recent layout, handwriting rigorously identical across twelve months of archives, sheets too clean to have come out of a mock-up session.
An honest reconstruction, on the other hand, is perfectly acceptable, and it is the only workable path. It carries today’s date, today’s signature, and explicitly states what it is: a reconstruction. It lists the elements that corroborate it — the day’s schedule, the centre’s entry log, the mock-up or fire-pit booking, the fire extinguisher recharge invoice, the instructor’s logbook — and states its degree of certainty. It declares itself; it does not hide. An operator that presents a documented reconstruction, a root-cause analysis, and a corrective plan is not treated the same as one where a fresh signature under an old date is discovered: the first has a process discrepancy, the second has a culture problem.
And when nothing corroborates it, the simplest solution remains, the one that always gets dismissed too late: redo the exercise. A session costs on the order of half a day of availability per person. A major finding on crew qualification costs something else entirely, and is paid for twice — once in correction, once in stepped-up oversight.
What to Go and Check in Your Own Documentation
Three points must be read in your own manual, not in a general guide, and they can be read in an hour.

- What is the reference record? Paper or system? Many manuals let both coexist without settling the question. The day they diverge — and they do diverge — it is the auditor who decides which one prevails, and they will choose the one that puts the operator at fault. Write down which of the two takes precedence, and align practice accordingly, including destroying the secondary medium.
- What is your electronic validation worth? A checkbox is only an attestation if the tool can state who checked it, when, and what was modified afterward. An editable field without a change log proves nothing at all; worse, it creates a risk of its own, since it allows the silent rewriting of the past and makes the entire database contestable with a single argument.
- Who holds what, and for how long? The retention period, the fate of the file after a crew member leaves, the crew member’s right to obtain their own history: these three points fall under the manual and the contract, rarely both in a coherent way. The practical consequence is seen especially in seasonal operations and in agency-supplied work, where the same crew member passes through several employers within a few years: whoever leaves without an extract of their own history arrives at the next employer unable to demonstrate anything, and the next employer redoes everything — that is, pays a second time for training that did take place. Providing in the manual for the handover of a dated, signed individual extract on the day of departure costs nothing and eliminates the handover dispute before it arises.
One grey area remains that no one has cleanly settled: remote-learning modules. They produce an abundant digital trace — login, duration, score, timestamp — and weak proof of identity. As long as the manual does not state which parts of these modules count as presence and which count as execution verified by a third party, the question will reopen at every audit, and the answer will depend on the auditor rather than the text.
For anyone preparing an instructor qualification, the subject changes in nature. From the day you sign, your own file becomes a piece of everyone else’s file: your validity, your standardisation, and your dates become enforceable against fifty people who are unaware of it. This is the shift that the FL330 training programmes treat as a chapter in its own right, because it takes almost every new instructor by surprise: you stop being a line in a register and become its source.
The applicable regulation, the means of compliance published by the authority, and your operator’s approved documentation prevail in all circumstances over this article, which describes a mechanism and not a procedure; in case of divergence, it is your Operations Manual that is authoritative.
Questions fréquentes
Is a collective attendance sheet enough to prove a practical exercise?
No. An attendance record proves that a person was present at a session, not that they completed each of the exercises planned that day. As soon as a workshop is postponed, split, or a person arrives late or leaves early, the sheet claims more than what actually happened. The evidence that holds up is a validation per exercise and per name, recorded on the spot at the moment the action is witnessed.
What should be done when the date of an exercise cannot be found in the file?
You reconstruct it, without ever backdating. The reconstruction carries today’s date, today’s signature, and explicitly states what it is; it lists the elements that corroborate it — the day’s schedule, the centre’s entry log, the mock-up booking, the instructor’s logbook. If nothing corroborates it, the only clean way out is to redo the exercise, which costs on the order of half a day and closes the matter for good.
Is an electronic validation acceptable in a cabin crew training file?
It is, if the tool can state who validated it, when, and what was modified afterward. A checkable box in an editable field without a change log is not an attestation: it allows the silent rewriting of the past and makes the entire database contestable. The manual must designate the reference record, paper or system, and practice must align with it unambiguously.
Who is responsible for the file when training is subcontracted?
The operator, under all circumstances: the demonstration of its own personnel’s qualification cannot be delegated. The provider most often sends back an end-of-course attestation, that is, a summary, while the detailed exercise-by-exercise record stays in its own files. The training contract must name that detailed record as a deliverable, with dates per exercise and the identity of the instructors, otherwise the granularity disappears the day the operator changes provider.
Does the expiry of an instructor's qualification call into question the sessions they signed?
Yes, for sessions signed after the expiry date, and the defect is never isolated: it falls in batches of files. The exercise may have been conducted perfectly, but the attestation was given by someone the operator no longer authorised to give it. This is why the instructor’s validity on the date of signature is an integral part of the evidence, just as much as the date of the exercise itself.
How can a file audit be prepared without spending weeks on it?
By pulling a single thread, in both directions. You start with a name picked at random from the system extraction and trace it back to the original document, then you start from a session sheet and check that each name on it can be found, exercise by exercise, in the individual files. Half a day per fleet, twice a year, reveals what a checklist never sees, because a checklist verifies that procedures exist, not the continuity of facts.

Leave a Reply
You must be logged in to post a comment.