Mustafa Habiyaremye

I spent four years in Air Operations for the United Nations in Mali. Peacekeeping air operations are an unusually demanding place to learn the job: the weather products are thinner than you would like, the airfields are not always what the chart says, and the consequences of a sloppy plan are not abstract. It teaches you quickly that operational control is a discipline, not a job title.

Then the mission closed and the job disappeared without warning. That is a normal enough story in aviation, and it is part of why the material on this site is written the way it is — for people who need to genuinely understand the work rather than pass through it.

What I do now is teach: video lessons on The Flight Dispatch Room, the written lessons and worked calculations here, and the workbooks and study guides in the shop.

Aviation knowledge is not scarce. Good explanation of it is. Most candidates preparing for the FAA ATP or Aircraft Dispatcher examinations are not short of material — they are short of material that shows the reasoning, separates the regulation from the interpretation, and is honest about which details actually decide an answer.

The lessons on this site are written to close that gap. Every one carries the worked calculation, the specific misreadings that produce wrong answers, and the official reference so nothing has to be taken on trust.

AVIA Flow started from the questions that kept arriving after each video on The Flight Dispatch Room — the same handful of topics, asked in the same confused way, by people who had already read the handbook. Video answers a question once. A written lesson with worked examples, practice questions and references answers it permanently.

The second reason is AI. Aviation professionals are already using it, whether or not their organisation has a policy on it. There is very little guidance written by someone who understands both what these tools are useful for and what operational control actually requires. That gap is worth filling carefully.

AI is genuinely useful in aviation work. It structures notes, drafts documents, generates practice questions, and organises information faster than doing it by hand. Every one of those is a task where being wrong is recoverable.

It is not useful for supplying an operational value, and it never carries authority. A model does not know the current TAF, your operations specifications, your MEL, or what has changed since the release was signed. It produces confident text either way, and confidence is not a signal of accuracy.

So every workflow on this site is built the same way: AI does the structuring, you supply the source material, and the verification step is written into the workflow rather than left implied. That is the only version of this that belongs anywhere near an operation.