SkyOwl
Book a walkthrough

Every training record holds a lesson. We find it

Observation, flight data, SOPs and standards: SkyOwl follows evidence to root cause, and shows you what to train next.

Who SkyOwl is for

  • Instructors

    Say what you saw. The debrief and grades come back drafted, with evidence, for you to finalise.

  • Students

    Read a debrief that makes sense: every point backed by your own flight, with practice that fits you.

  • Training managers & heads of training

    See where each cohort is weak, and generate interventions at scale.

  • Human factors compliance

    Gather the data to identify human factors training gaps and comply with HF / NTS requirements.

All flights are logged. Few are learned from

Debriefs, grades and flight data pile up faster than anyone can read them. The patterns that would change the training stay hidden.

Five pressures on every training organisation:

  • Thousands of new pilots
  • Inconsistent grading
  • Instructor turnover
  • Type conversions
  • New ways of learning

Reason, learn, classify, intervene

One engine reads everything a training organisation already produces, and closes the loop with the right exercise for the right person.

  1. Reason

    Instructor notes, manuals, incident reports and flight data, read together instead of in separate silos.

  2. Learn

    Our model cross-references data from multiple sources and follows the evidence to a root cause. If it can’t point to the evidence, it doesn’t say it.

  3. Classify

    Across every session, results sort themselves by competency. Trends surface with no manual coding.

  4. Intervene

    Each gap arrives with something to practise — a quiz, a scenario or a short game — for the person who needs it.

The debrief writes itself

In the language of your competency framework, non-technical skills included. Every sentence names the behaviour it saw — pick a number to see where it is filed.

Debrief · 787 recurrent, EBTDraft · for the instructor to finalise

Overall

You flew a recurrent session with good crew coordination and decision-making, but two separate lapses brought the outcome below standard. You ran the checklist from memory before confirming with the QRH, which was poor procedural discipline. During the hand-flown SID, bank reached 32 degrees in the turn and you did not correct until the PM called it. These are unrelated errors — one traces to incomplete checklist protocol, the other to a monitoring lapse during manual flight — so each needs separate attention.

What went well

The pairing brief was thorough. You requested the storm deviation early, and the PM set out the avoidance plan clearly. The PM briefed the cabin without being asked. The diversion decision was reached quickly, both pilots contributed, and fuel figures were cross-checked twice.

What the instructor saw backed by the flight

  1. Observed

    SkyOwl’s notes screen: the flight recording has finished, and the instructor’s note reads “inability to use ailerons vs rudder effectively”.

    The instructor’s note, typed or dictated during the sortie, beside the flight recording.

  2. Measured

    +718 ft above the nominated height, against CASA Part 61 MOS A5.3.2.

  3. Explained

    Likely causes, for the instructor to confirm, then the next lesson to practise.

Trends, made insightful

Patterns across every debrief surface on their own, and each gap arrives with something to practise, fitted to the trainee who needs it.

Below standard, by competency · illustrative

  1. Decision making
  2. Workload
  3. Procedures
  4. Flight path
  5. Knowledge
  6. Communication

Top gap: decision making

Six years of research. Tested with real pilots and instructors

  • 2,700real pilot interactions
  • 150+hours of expert interviews
  • 5,000incident reports indexed
  • 150,000knowledge-graph documents
  • 74%agreement with instructors across 1,700 grades

Built and tested with

  • RMIT Aviation Academy
  • Air Chathams
  • Australian Government: Australia’s Economic Accelerator
  • Pleias

Pedagogy, aviation and AI — in one lab

SkyOwl grew out of six years of research at the University of Melbourne, under the Southern Cross.

The Southern Cross

  • Fabio Mattioli

    Founder

    Associate Professor, University of Melbourne. AI for human factors and training.

  • Natalee Johnston

    Domain lead

    Former Royal Australian Navy pilot and instructor.

  • Hiruni Kegalle, PhD

    Data & adoption

    Builds the data layer, and the research that makes trainers trust it.

  • Kerry Phillips

    Feedback & resources

    Flight instructor. Develops pedagogical resources that work in practice.

  • Tristonne Forbes

    Growth & strategy

    20 years leading innovation programmes.

Start with one cohort. Grow with your data

Every step runs on the same engine and adds more of your data. You see results on your own records before you take the next one.

Book a 30-minute walkthrough

Or write to info@skyowl.tech

  1. Start here

    Consistent debriefs

    Notes graded against your standard, with evidence, in reports every instructor writes the same way.

  2. Then

    Training that fits each pilot

    Each trainee’s root causes become the right exercise between sessions, and the cohort’s trends show what to train next.

  3. Where it goes

    Instructor time where judgement matters

    Connect simulator and flight data to your records: what can be measured is measured, and instructors spend their time on what only they can judge.