Every serious operation has a FRAT. Few have a FRAT that works.
The difference is not in the form — it is in when and how it is filled out. The FRAT that was born to structure the decision to fly becomes, over time, a ritual: filled out in a hurry in the crew room, from memory, after the decision to fly has already been made. The paper says "we assessed the risk"; the practice says "we documented the decision that was already made."
If your FRAT has never changed anything — never delayed a departure, never escalated a decision, never added a mitigation — it is not measuring risk. It is measuring the patience of whoever fills it out.
PAVE is good precisely because it is tedious
The PAVE framework divides flight threats into four families:
- P — Pilot: the pilot. Rest, recent experience on type, proficiency in the type of operation of the day, personal pressure.
- A — Aircraft: the aircraft. Open discrepancies, upcoming expirations, inoperative equipment and what it changes in the flight's capability.
- V — enVironment: the environment. Weather along the route and at the airports, runway conditions, NOTAM, night operation, terrain.
- E — External pressures: external pressures. The passenger who needs to arrive, the slot window, the cost of cancellation, the "it has always worked out."
The strength of the method lies in forcing you to look at each family separately. The accident rarely comes from one giant, obvious threat; it comes from the stacking of medium threats — a tired pilot, in an airplane with an MEL item, bound for a short, wet runway, with a passenger in a hurry. Each one, on its own, would pass. PAVE exists so that the stacking shows up before takeoff.
And that is why it is laborious: truly assessing the four families requires gathering information that lives in different places — the METAR/TAF on one site, the NOTAM on another, the maintenance status in the maintenance tracking, the crew's rest in the roster. The cost of collection is the real reason for so many FRATs filled out from memory.
Where the assessment fails in practice
Failure in collection, not in judgment. The captain knows how to judge a wet runway. What is missing at 6 a.m. is having at hand the runway condition, the forecast wind in the landing window and the NOTAM that closed the taxiway — without opening five sources.
A score without consequence. A score that triggers nothing is decoration. A mature FRAT defines bands: up to here the crew decides; above this, mitigate and document; above that, the decision goes up a level — before the flight, with a named decision authority.
A threat without mitigation. Identifying "strong crosswind" and stopping there is halfway. The question that closes the assessment is: what changes in the flight because of this? A revised alternate, extra fuel, a specific approach briefing — mitigation is what turns the list into a plan.
The record that gets lost. A paper FRAT dies the next day. The historical series — which threats show up every week, which route concentrates a high score — is direct input to the SMS: it is your hazard map being drawn flight by flight, for free.
What AI can do here — and what it cannot
AI fits well exactly at the step that blocks everything: collection and organization. Reading the airport's ROTAER, crossing route weather with the flight window, scanning NOTAM, gathering the aircraft's status — and returning it already organized into the PAVE families, with proposed severity and suggested mitigation for each threat.
What it does not do — and should not do — is decide. The final severity, the acceptance of risk and the decision to fly are the judgment of whoever operates, with technical responsibility that is not delegated to a model. AI that promises to "approve" the flight is solving the wrong problem: the bottleneck was never pressing the button — it was seeing the complete picture in time to decide well.
How Alis handles this
Alis's Flight Kit generates the risk assessment as part of the pre-flight briefing, and AI works on the collection:
- PAVE threat inventory, assembled by AI: the platform reads the ROTAER and the official sources for the airports on the route — origin, destination and alternates —, crosses them with weather, route and winds, and returns the threats classified by family and severity.
- Mitigations and briefings per threat: each item comes with a proposed mitigation and goes into the flight's operational briefing — not onto a separate sheet.
- Explicit decision authority: the assessment indicates the recommended decision level, so the conversation happens before takeoff, not after.
- A record that becomes history: each FRAT is linked to the flight, the aircraft and the crew — searchable, auditable and ready to feed the SMS.
The captain remains the owner of the decision. The difference is that he or she decides looking at the whole picture — assembled in minutes, not in five browser tabs.
If your operation is structuring its risk management, it is also worth reading how to move out of the filing cabinet in the SMS and getting to know the electronic logbook that feeds this data at the source.
This article is informational and does not replace current ANAC regulations, your operation's approved manuals or the operational judgment of whoever holds responsibility for the flight decision.