AI for aviation and airlines
AI you can audit. For airline operations, training and safety-critical systems.
SkyDynamics builds AeroBrain, a set of auditable AI technologies for safety and mission-critical work, and the products that use it: decision support for airline operations control, an AI training tool for EBT and CBTA, and real-time data on the aircraft. A person decides; every answer keeps its evidence.
What AI for aviation takes
Verifiable, or not used. Aviation needs AI whose answers can be checked.
AI for aviation has to be verifiable. In AeroBrain retrieval is deterministic, every answer keeps its evidence with its source and revision, and the human decision is recorded with it. The software is developed to safety-critical rules such as MISRA C and the objectives of DO-178C. Developing to a standard is not a certificate, and the site does not claim one.
- A general assistant
- An answer
- AeroBrain
- An answer with its evidence
Evidence with every answer
The input data and its time, the evidence with its source and revision, the model version, the output and the human decision are recorded together.
Replayable
Retrieval is deterministic, so an auditor can replay the same question on the same data and obtain the same evidence.
A person decides
AI proposes and shows its reasons. The controller, the instructor or the engineer decides, and regulated actions require four-eyes approval.
From the AeroBrain page, where the technologies and the standards they are developed to are set out in full.
Where it works today
Four uses. Operations, training, maintenance training and the pilot.
AI for airline operations: OpsEye
Disruption management and operations control built on AeroBrain. When a flight is disrupted it proposes ranked, costed recovery plans shaped by the airline's own policy and KPIs, with crew duty limits, the MEL, curfews and minimum connection times checked before a plan is shown. The duty manager decides.
Learn moreAI training tool: the EBT Scenario Builder and AeroEBT
AeroBrain drafts EBT scenarios from the operator's safety data and aircraft procedures, and checks topic coverage and difficulty while the designer writes. In the session it proposes observations and competencies from the instructor's notes. The instructor assesses; the AI assists and never steers.
Learn moreAI in maintenance training
In CBTA for Part-147 and Part-145 training, AeroBrain reads the assessor's comments on a task and suggests the observable behaviours they evidence. The assessor confirms each one.
Learn moreThe technology: AeroBrain
A secured RAG engine written in Rust, the Contextual Composition Protocol, the LLiDA real-time database and an auditable Assurance Small Language Model, for real-time decision support from avionics data and approved manuals.
Learn moreReal-time data: LLiDA
A real-time time-series database that works in RAM, on disk and at processor level and keeps every node synchronised, from the aircraft to the ground.
Learn moreFor the pilot: Flying Whale
The SkyDynamics app for pilots: briefing aids, weather, NOTAMs and a logbook on the iPhone.
Learn more
Safety-critical and mission-critical AI
Built for the cases where a wrong answer costs. In aerospace and defence.
Safety-critical AI supports work where an error can harm people or an aircraft. Mission-critical AI supports work that must not stop. AeroBrain is designed for both, with these properties.
Deterministic retrieval
The same question on the same data returns the same evidence, so a result can be reproduced and audited.
Real time by design
Built to answer while the event is still unfolding. For hard real-time functions each stage is designed to run inside a fixed time budget, set with the customer and shown for each product.
Developed to flight-software rules
Written to safety-critical rules such as MISRA C and the objectives of DO-178C. This is how the software is developed; it is not a certification.
Human decision on record
The human decision is stored with the answer and its evidence. AeroBrain never holds an operational decision.
On the edge and offline
A compact model and the LLiDA database run on the device, so the function does not depend on a connection to a server.
Classified by its level of assistance
SkyDynamics classifies AeroBrain as EASA AI Level 1B, human cognitive assistance, with reference to the EASA AI Concept Paper. That is our own classification, not an EASA approval.
By use
One AI layer. A named person accountable in each use.
| Criterion | What the AI does | Who decides | Product |
|---|---|---|---|
| Airline disruption | Ranks and costs recovery plans within the airline's rules | The duty manager | OpsEye |
| EBT scenario design | Drafts the scenario and checks coverage and difficulty | The training team | EBT Scenario Builder |
| Simulator session | Proposes observations and competencies from the notes | The instructor | AeroEBT |
| Maintenance task assessment | Suggests the behaviours a comment evidences | The assessor | AeroEBT for maintenance training |
| FSTD qualification | Points out gaps and missing evidence | The operator and the authority | FCS++ |
In every row the AI assists and a person decides. Nothing on this page describes an autonomous system.
AI for aviation: questions people ask
What kind of AI is suitable for aviation?
AI for aviation has to be verifiable. AeroBrain is built for that: retrieval is deterministic, every answer keeps its evidence with its source and revision, the human decision is recorded with it, and the software is developed to safety-critical rules such as MISRA C and the objectives of DO-178C. It is designed to give decision support in real time from avionics data and approved manuals, for safety and mission-critical work in aerospace and defence.
Which AI is used by airlines for operations control?
OpsEye, from SkyDynamics, is an airline disruption management and operations control system built on AeroBrain. It gives a disruption index per flight with its contributing factors and proposes ranked, costed recovery plans shaped by the airline's policy and KPIs. A person decides, and regulated actions require four-eyes approval.
Is there an AI training tool for pilot training?
Yes. The EBT Scenario Builder and AeroEBT use AeroBrain as an AI training tool for evidence-based training (EBT) and competency-based training and assessment (CBTA): it drafts scenarios from the operator's safety data, checks topic coverage and difficulty, and proposes observations from the instructor's notes. The instructor remains the assessor.
What is safety-critical AI?
AI that supports work where an error can harm people or an aircraft, so its answers must be traceable and reproducible. AeroBrain is designed for it: deterministic retrieval, evidence kept with each answer, the human decision on record, and software developed to rules such as MISRA C and the objectives of DO-178C.
What is mission-critical AI?
AI that supports work that must not stop, such as operations control or defence missions. AeroBrain is designed to keep working on the device, in real time and without a connection to a server, with a compact language model and the LLiDA real-time database.
Is AeroBrain certified?
No certificate is claimed. AeroBrain is developed to safety-critical rules such as MISRA C and the objectives of DO-178C, and SkyDynamics classifies it as EASA AI Level 1B, human cognitive assistance. Developing to a standard is not a certification, and the classification is our own, not an EASA approval.
Does the AI make the decision?
No. In every SkyDynamics product the AI proposes and shows its reasons, and a person decides: the duty manager in operations control, the instructor in a training session, the assessor in maintenance training.
See it on your own case.
Tell us the decision you want supported: a disruption, a scenario, an assessment. We show how the evidence is kept and who decides.
We respond within one business day.