Agentic AI for Flight Training Intelligence

This demo shows how an AI agent can ingest data from Aerosim's Console 40 simulator sessions, iFITS gaze-tracking, and CloudAhoy flight analytics — then generate personalised training recommendations with instructor approval before any action is taken.

42%

Time Saved

Students

24/7

Monitoring

Avg. Training Efficiency

78.4%

+12.3% vs. manual tracking

AI Interventions This Week

5 approved, 3 pending

Predicted Time-to-CPL

8.2 mo

-1.4 mo vs. cohort avg

Gaze SA Score (Cohort)

72/100

+8 pts since week 1

Active Student Pilots — Cohort 2026-A

Student	Program	Sim Hours	Gaze SA	Flight Score	Status	AI Flag
Alex Chan	PPL	47.2h	85	82%	On Track	None
Sarah Wong	CPL	112.8h	91	88%	Excelling	None
Mattis Tsang	PPL	38.5h	52	61%	At Risk	Intervention Needed
Emily Lee	PPL	29.1h	68	74%	Monitor	Gaze Pattern
Marcus Tan	CPL	95.3h	78	79%	On Track	None
Priya Patel	PPL	18.7h	61	69%	New	Baseline
Daniel Garcia	CPL	128.4h	88	91%	Check Ride Ready	None
Yuki Nakamura	PPL	42.6h	58	72%	Monitor	SA Regression

Agentic AI Analysis — Mattis Tsang (PPL)

Ingesting Simulator Data

Pulling 47 sessions from Console 40 via CloudAhoy API...

Fetching: cloudahoy.com/api/v3/flights?student=mattis_tsang Sessions: 47 total (38.5 flight hours) Aircraft: DA-40 (G1000 glass cockpit) Date range: 2025-11-15 to 2026-02-28

Analysing iFITS Gaze-Tracking Data

Processing eye-tracking patterns for situational awareness scoring...

iFITS Gaze Analysis for MATTIS_TSANG: SA Score: 52/100 (Below Threshold: 65) Primary Fixation: PFD 62% | MFD 28% | OTW 7% | Other 3% Critical Finding: Outside-the-window scan rate 73% below cohort avg Pattern: Fixation tunneling on airspeed indicator during approach phase (sessions 38-47) Trend: SA declining since session 31 (-18 pts)

Cross-referencing FAA Part 61 Requirements

Mapping performance gaps to regulatory training standards...

FAA Part 61.107 - PPL Aeronautical Experience: [PASS] 61.107(b)(1) - Normal procedures [WARN] 61.107(b)(1)(iii) - Airport/traffic pattern ops (68%) [FAIL] 61.107(b)(1)(iv) - Instrument ref. flight (54%) [WARN] 61.107(b)(1)(v) - Navigation (65%) [PASS] 61.107(b)(1)(ix) - Night operations (71%) Risk: Student at risk of failing instrument reference component of practical test at current trajectory.

Generating Personalised Training Plan

Building adaptive curriculum with prioritised focus areas...

PERSONALISED TRAINING PLAN - Mattis Tsang Generated: 2026-03-02 | Confidence: 87% Priority 1 (URGENT): Instrument Reference Flying Action: 4x dedicated IFR hood sessions (Console 40) Focus: Cross-check scan pattern, reduce PFD fixation iFITS: Enable real-time gaze feedback alerts Target: SA score > 65 within 6 sessions Priority 2: Traffic Pattern Awareness Action: 2x pattern work with emphasis on OTW scanning Focus: Wind correction, spacing, OTW lookout discipline Target: Airport ops score > 75% Priority 3: Navigation Skills Action: 1x cross-country planning + 1x sim session Focus: Pilotage + dead reckoning, VOR tracking Target: Navigation score > 72% Estimated impact: +3.2 weeks to PPL if addressed now +8.7 weeks if unaddressed (risk of plateau) STATUS: PENDING INSTRUCTOR APPROVAL

Student Profile

Mattis Tsang

PPL Program • Enrolled Nov 2025

At Risk

Sim Hours

38.5h

Sessions

Gaze SA

52/100

Est. PPL

+3.2 wks

Key Findings

Fixation Tunneling Detected

During approach phases, student fixates on airspeed indicator for 8-12 second intervals, missing altitude and course deviation checks. This pattern emerged at session 31 and is worsening.

Low Outside-the-Window Scan

Only 7% OTW fixation time vs. cohort average of 18%. This impacts traffic awareness and visual approach skills.

Instrument Reference Below Standard

54% score on FAA 61.107(b)(1)(iv) — below the 70% threshold for practical test readiness. Correlated with gaze pattern issues.

Instructor Action Required

The AI agent has generated a training intervention plan. Per safety gate protocol, instructor approval is required before changes are applied to the student's curriculum.

Autonomous Orchestrator — Closed-Loop EVAL/ACTION

Idle

Objective

Autonomy Aggressiveness

70% autonomy aggressiveness

Safety Gate

Instructor Approval Required

Actions proposed by AI, executed only after approval

Tool Trace

Awaiting orchestrator run...

AI Ranked Intervention Queue

No run yet

Pending

Run the loop to evaluate risk, draft actions, and re-evaluate residual risk.

Loop Closure

EVAL → ACTION → RE-EVAL has not run yet.

Agent Memory

No prior runs in this session.

Cockpit Gaze Distribution — Mattis Tsang (Last 10 Sessions)

62%

PFD (Primary Flight Display)

28%

MFD (Multi-Function Display)

Outside-the-Window

Other

AI Alert: Abnormal fixation pattern detected

PFD fixation at 62% is significantly above the healthy range (35-45%). Student shows "tunnel vision" on airspeed indicator, particularly during approach and landing phases. Recommend iFITS real-time gaze feedback alerts during next 4 sessions.

Cohort Benchmark — Healthy Scan Pattern

40%

PFD (Primary Flight Display)

25%

MFD (Multi-Function Display)

18%

Outside-the-Window

17%

Engine / Switches

Healthy Pattern: Balanced instrument cross-check

A well-trained pilot distributes attention across all instruments and the outside environment. The 18% OTW rate is critical for traffic awareness and visual approach skills. This benchmark is derived from 240+ student sessions on Console 40.

SA Score Trend — Mattis Tsang (47 Sessions)

Session 1 Regression starts → Session 47

Above threshold (65+) Warning zone (55-64) Below threshold (<55)

Pending AI Recommendations — Instructor Approval Required

3 Pending

Urgent Training Intervention — Mattis Tsang

AI Agent detected fixation tunneling and declining SA scores (52/100, down 18pts). Recommends 4x dedicated IFR sessions with iFITS real-time gaze alerts enabled. Estimated impact: prevent 3.2 week delay to PPL.

Generated 2 hours ago • Confidence: 87%

Moderate Gaze Pattern Alert — Emily Lee

SA score at 68 with inconsistent scan pattern during cross-wind approach scenarios. Recommends enabling iFITS visual scan prompts for next 3 sessions. Low risk if deferred 1 week.

Generated 5 hours ago • Confidence: 74%

Moderate SA Regression Warning — Yuki Nakamura

SA score dropped from 72 to 58 over last 8 sessions. Pattern correlates with transition from VFR to IFR training phase. Recommends 2x bridge sessions combining visual and instrument reference. May benefit from iAALP AR approach visualization before next sim session.

Generated 1 day ago • Confidence: 81%

Recently Approved

5 Approved

Approved Check Ride Readiness — Daniel Garcia

AI analysis confirms student meets all FAA Part 61 requirements for PPL practical test. Recommended scheduling check ride within 2 weeks.

Approved by Capt. Dean Shing • 2 days ago

Executed

Approved Accelerated Curriculum — Sarah Wong

Student exceeding CPL benchmarks. AI recommends advancing to multi-engine training 2 weeks ahead of schedule.

Approved by Capt. Dean Shing • 4 days ago

Executed

System Architecture — How AeroIntel Integrates with Aerosim's Ecosystem

Data Sources

Console 40
DA-40 Simulator

iFITS
Gaze Tracking

CloudAhoy
Flight Analytics

ForeFlight
EFB Data

iAALP
AR Learning

↓

Agentic AI Engine

Data Ingestion
Multi-source orchestration

→

Performance Analysis
ML + regulatory mapping

→

Plan Generation
Personalised training paths

↓

Human-in-the-Loop Safety Gate

Instructor Review & Approval
No action taken without human authorization

↓

Outputs

Training Plans
Adaptive curriculum

Student Reports
Performance analytics

Compliance Docs
FAA/CASA/HKCAD

Alerts
Early intervention

Impact: Manual Process vs. AeroIntel Automation

Before: Manual Process

~4.5 hours per student per week

Instructor reviews sim recording

45 min per session

Manually check gaze-tracking logs

30 min, prone to oversight

Cross-reference with FAA standards

20 min, requires deep knowledge

Write student progress report

30 min per student

Adjust training plan manually

15 min, based on intuition

→

After: AeroIntel Automation

~15 min per student per week

Auto-ingest all session data

Instant via CloudAhoy + iFITS API

AI analyses gaze + flight data

30 seconds, catches all patterns

Auto-maps to FAA/CASA requirements

Instant, always up-to-date

Generates report + training plan

AI-generated, instructor reviews

Instructor approves in 1 click

15 min review, data-driven decisions