Descripció del projecte
Context:
Infotainment has become one of the most complex and customer-visible systems in modern software-driven vehicles (and in the medium term in software-defined vehicles), combining HMI, applications, connectivity, vehicle signals, market variants and countless user scenarios. This complexity makes traditional validation increasingly demanding in terms of effort, coverage and speed. Generative AI, AI agents, advanced validation automation and machine learning now create a unique opportunity to transform testing into a smarter, scalable and continuously improving process.
Objective:
The project aims to develop an AI-assisted validation ecosystem for CUPRA automotive information: an intelligent “superagent” capable of connecting existing tools, executing end-to-end tests, analyzing system behavior, learning from the results and dynamically improving test coverage, without lacking reliability.
The framework will combine automation, AI agents, GenAI, data engineering, real testbeds, vehicle environments, digital twins, documentation, historical defects, customer complaints and human feedback in a loop. The ambition is to move from static and mostly manual validation towards a proactive, adaptive and data-driven approach.
Through intelligent agents, the framework will be:
– Execute and improve Infotainment testing cases autonomously
– Interact with HMI screens, signals and components of vehicles, buses, test systems and validation tools
– Detect unexpected behaviors beyond predefined specifications, scenarios and test cases
– Pre-analyze failed tests and autonomously generate emission reports using VW Group reporting systems
– Learn from human feedback to improve reliability
– Propose improved testing of documentation, defects, field claims, customer feedback and multiple sources
– Create automated status reports based on test results and their analysis, as well as additional internal feedback
– Explore optimization of computational resource usage, as well as hybrid cloud approach to enable local execution in real time
The project will contribute to:
– Greater validation efficiency and wider coverage
– Early detection of customer impact issues and forecasting of future problems
– Stronger internal AI competition
– Better test prioritization based on real data
– More robust launch recommendations
– Improving the customer experience in a key domain of the digital vehicle
