Ask a ground handling operation whether Anna is qualified to marshal an aircraft and you’ll get an answer in about a minute. Someone opens a spreadsheet, finds her row, and reads across.
Ask whether Anna was qualified at 06:40 on the fourteenth of March, and the room goes quiet.
That’s the whole problem in one question, and — as the new EU ground handling rules make clear — it’s about to matter a great deal more than it does today.
Current state is not a record
The spreadsheet answers the present tense. Is this person qualified, right now, according to the last time someone updated the file.
It’s a perfectly reasonable thing to have. It’s also structurally incapable of answering a question about the past, because when Anna renewed her certification in April, someone typed over the old expiry date.
The previous state didn’t get archived. It got destroyed.
So the file tells you Anna is qualified today. It cannot tell you what it said in March. And “what did it say in March” is exactly what you’ll be asked, either by an authority under the new rules or by an insurer after something goes wrong.
The fix isn’t a better spreadsheet. It’s a different shape of data. Records about people need to be append-only — you add a new fact with a date, you never overwrite the old one. Then “was Anna qualified on 14 March” is a query rather than an act of memory.
This sounds like a technical detail. It’s the difference between having evidence and having a claim.
What actually needs tracking
More than most operations realise, and each item has its own clock.
Six clocks, per person
- The certification itself, with the date granted and the date it expires.
- The recurrent training cycle, usually on a different period than the certification.
- The on-the-job component — who observed, when, against what standard, and whether they signed.
- The airside driving permit, with its own renewal and its own issuing body.
- Any medical or licence requirement attached to specific equipment.
- The human factors and emergency response training the regulation now expects.
Six clocks, running independently, per person. On a shift of forty that’s two hundred and forty expiry dates, and there is no version of that which a person tracks reliably by looking at a file once a month.
What happens in practice is that expiries are discovered rather than anticipated. Someone notices at the start of a shift that a permit lapsed last week. Now you’re short a qualified person with no notice, on a morning that had no slack in it.
The part everyone misses: a qualification is a scheduling constraint
This is the argument I care most about, because it changes what kind of problem this is.
Qualification data is normally owned by training or HR. It sits in their system, on their schedule, reviewed on their cycle. The rostering system, meanwhile, allocates people to tasks — and in a lot of operations it either doesn’t know about qualifications at all or holds a copy that goes stale.
So the planner produces assignments that people aren’t permitted to perform. The supervisor catches most of them, because the supervisor knows their people. Then the supervisor is off, or there’s a swap at 05:50 to cover a no-show, and one gets through.
You now have a task performed by someone not certified for it, and — this is the bit that turns a near miss into a problem — no record either way. You can’t prove it didn’t happen and you can’t prove it did.
Feed qualifications into the planner as hard constraints and the assignment simply cannot be generated. Not “flagged for review.” Cannot be generated. The person doesn’t appear as a candidate for that task, in the same way that someone who’s already working can’t be in two places.
And the record of who was assigned to what, under which qualification, valid until when, falls out of planning as a by-product. You didn’t build an audit trail. You built a roster, and the audit trail is what a roster is when the data underneath it is honest.
The reverse direction, which is where the money is
Once expiry dates and the roster live in the same place, something more useful becomes possible, and almost nobody does it.
Recurrent training has to happen. It takes people off the line for a day. In most operations it’s scheduled when the training department has capacity, or when someone notices an expiry approaching — which means it lands wherever it lands, including in the middle of a peak week when you can least afford to lose a qualified person.
But you know the flight schedule. You know the demand curve. And you know, months ahead, exactly whose certification expires when.
So you can put the training in the troughs. Deliberately. Move the sessions to the quiet weeks, batch the people whose expiries fall in the same quarter, and stop paying overtime to cover a classroom.
You can also see the cliff coming. If eleven people’s certifications expire in the same six weeks — which happens constantly, because they were all hired or certified together — you find out in advance rather than in the week it bites.
That’s a workforce planning capability you get almost free once the data is in one place and has dates on it. I’d argue it pays for the whole exercise before compliance is even mentioned.
What I’d fix, in order
Make the record append-only. Stop overwriting. Every qualification event gets a date and stays. This is the change that turns a spreadsheet into evidence, and it costs nothing but discipline.
Put every clock in one place. Certification, recurrent, on-the-job sign-off, driving permit, medical. Six systems means six things nobody checks.
Alert on lead time, not on the date. An expiry warning on the day it expires is a staffing emergency. Ninety days out, it’s a training booking. Same information, entirely different cost.
Let the planner read it. This is the step that converts an administrative asset into an operational one. Until the rostering system treats qualification as a hard constraint, you’re relying on supervisors to catch what the software generates — and the morning roster is already being overwritten more often than anyone counts.
Then plan training against demand. Once the first four are done this one is nearly free, and it’s the one that shows up in the budget.
Notice that only the first item is about compliance. The rest are things worth doing if the EU had never written a regulation. That’s usually a good sign that you’re solving the real problem rather than the paperwork.
Optimum Intelligence builds operational systems for ground handling. OptiCheck holds people, skills and availability as structured data — which is what lets a roster respect qualifications rather than assume them. If your training matrix and your rostering system have never spoken to each other, we’re happy to talk.