We will discuss how AI can help structuring heterogenous legacy tests data in a consolidated data base (enabler to any AI/ML implementation). A lot of legacy data exists but heterogenous format prevent from implementing AI analysis on this data. • Test result analysis using ML: o Fault localization on avionics systems o No Go anticipation by identifying weak patterns during test program execution o Predictive maintenance, Remaining Useful Time computation • Open Neural Network Exchange integration in SIL/HIL systems • How ML can accelerate test program/test hardware design and validation