AI for continuing airworthiness
For the CAMO that wants the speed of AI and still answers for every result.
An AI in the CAMO can only be made ready by a CAMO with AI.
Continuing airworthiness runs on records, rules and responsibility. Our AI works through all three. It reads the records, applies the rules one narrow question at a time, cites its source for every result and leaves every decision to a qualified person.
Nothing reaches your records until a named person has approved it. That is what makes it usable in a CAMO.
One agent per parameter applies strict logic to one narrow question and cites its source. A final step combines the answers with minimal reasoning.
Where the risk calls for it, independent checks answer the same question. Only matching answers pass; disagreements go to an engineer.
A named person approves every step before anything reaches your records. Every change is logged per user.
Your data stays in your workspace under EU data protection and is never used to train models.
Four layers, each working on its own. Each one adds data, logic and write access to the one before it.
Layer 1
A regulatory copilot for individual users
Ask questions about regulations, AMC/GM and aviation law, and draft CAME sections and procedures. Every answer comes from a knowledge base that stays up to date for your region, and cites its source.
For
Individual users: engineers, auditors and managers.
Delivered
A closed copilot that works beside one person, not inside your IT landscape.
Status
Ready as an MVP.
Nothing. Layer 1 has no access to company or aircraft records.
Inputs
Processing
Outputs
Users
Layer 2
Records review for aircraft transitions
Review one aircraft’s records for compliance and receive the findings as comments in an Open Item List.
For
Lessors, operators and transition teams.
Delivered as product
Our first market product, for transitions. It changes no aircraft data.
Status
In operational tests, expected Q1 2027.
Comments only. Layer 2 does not change records.
Inputs
Processing
Outputs
Users
Layer 3
A system of record that evaluates and documents compliance
Decision pipelines work through CAMO tasks one parameter at a time. A person approves every step before anything reaches the database, which then drives work orders and task cards. User interfaces are flexible and state of the art.
For
CAMOs and operators.
Delivered as product
Software your team works in, with training and AI procedures written into your CAME. It runs alongside your current ERP during the transition.
Status
Trials and demo from Q2 2027. Release after EASA Part-AI is published.
Approved entries only, each logged against a named user.
Inputs
Processing
Outputs
Users
Layer 4
Fleet-wide engineering and planning from one database
The Operation layer across your whole fleet. Fleet tasks are generated from approved database entries and pass through the same pipelines and approvals.
For
Fleet and reliability engineering.
Delivered as product
Software your team works in. It builds on Layer 3.
Status
Expected Q2 2028. Depends on investment.
Fleet-wide entries, approved by a named user.
Inputs
Processing
Outputs
Users
Compliance
EASA’s NPA 2025-07 proposes a generic AI trustworthiness framework for aviation. Approved organisations such as Part-CAMO are within its scope.
Our design follows its main elements: risk assessment before use, human oversight, explainability, error management and in-service monitoring.
Sources: EASA, NPA 2025-07 · EASA, AI Concept Paper, Proposed Issue 3