This AI model was custom-trained on a dedicated dataset and runs locally on a mobile phone.
A custom-trained vision AI model detects three types of compressor
blades from a CFM engine with 95-99% confidence. ⭐⭐⭐⭐⭐
To preserve privacy, no cloud computing was used in this example. Larger, more accurate models can also run on a private cloud.
Can AI detect a missing bolt on a landing gear assembly?
Yes, with 95-99% confidence. ⭐⭐⭐⭐⭐
Landing gear inspection of an Avro Vulcan B2 aircraft.
This AI model was custom-trained on a dedicated dataset and runs
locally on an edge device (a mobile phone).
Can aircraft skin be inspected with a mobile phone?
Yes.
Aircraft skin inspection using an AI model for rivet detection. The custom-built model automatically detects, counts and tracks every rivet, assigns each one a unique ID and builds the inspection report.
Can AI detect a dent or scratch on a turbine disk or blisk?
Yes, with 95-99% confidence. ⭐⭐⭐⭐⭐
This is a challenging use case, because the apparent shape of each blade on a turbine disk or blisk changes with the viewing angle. The AI still detects anomalies and defects, and tracks and identifies each individual blade.
Our AI return on investment (ROI) tool helps you quantify and present the business case for implementing AI models such as those used in AI Visual MRO. It provides a quick assessment to support organisations evaluating the adoption of AI solutions. 04-May-2026
In our industry, missing or incorrectly fitted bolts can cause serious damage, even if such errors are rare. That is why it is good practice to have an AI system double-check the work after each assembly operation and record the results for future analysis. 04-May-2026
Detection of dents, scratches and other forms of foreign object damage (FOD) using AI. In this example, an AI model is deployed on an edge device (an Android device), enabling real-time visual inspection and alerting technical staff whenever an anomaly is detected. 06-May-2026
A comprehensive PowerPoint presentation covering all the information on this site, along with additional insights. Slide 10 includes an AI capability rating, from easy to difficult, giving a clear indication of the effort required to validate and deploy these AI models. 12-May-2026
AI Visual MRO is an advanced AI solution for visual inspection in maintenance, repair and overhaul (MRO) operations. AI models tailored to your needs automatically detect defects such as scratches, dents and corrosion, with accuracy typically between 87% and 99%, depending on the application.
Our AI models analyse images and video from virtually any camera or borescope to automatically identify and classify defects. Each model is trained on your own images, ensuring maximum accuracy for your specific use case.
AI Visual MRO delivers measurable results: 58% overall cost savings on visual inspection, 80% time savings on report generation, and 94% average accuracy in automatic defect detection.
Yes. These AI models can be trained on your own images and run locally on your infrastructure for maximum privacy and security. Your data never leaves your systems unless you choose to share it. This makes the solution ideal for the aerospace and defence sectors, where data analysis and AI models often need to run privately.
