Autonomous HMI Testing for a European Automotive OEM

The Challenge

A leading European automotive OEM was preparing to launch its next-generation infotainment platform across multiple vehicle lines. The system was a significant leap forward — featuring a redesigned digital cockpit, smartphone integration, connected services, and support for 38 language variants.

With over-the-air update cycles planned every 4–6 weeks post-launch, the OEM needed a testing approach that could keep pace. Their existing QA process — a combination of manual testing and scripted automation — was already struggling:

• Coverage gaps: The team could validate less than 15% of interaction paths per release cycle
• Script fragility: Each UI update broke 30–40% of existing test scripts, requiring weeks of rework
• Language testing: Verifying 38 language variants was practically impossible with manual QA — most received only spot checks
• Release delays: QA had become the primary bottleneck in the release pipeline

The Approach

Filuta AI deployed autonomous testing agents directly on the OEM's infotainment hardware. Rather than executing predefined test scripts, the agents used Composite AI — combining symbolic planning with machine learning — to autonomously explore the system.

Phase 1: System Modeling

Filuta's agents began by autonomously mapping the infotainment system — discovering screens, menus, controls, and transitions without any pre-built model or manual configuration. This produced a comprehensive system map that served as the foundation for systematic testing.

Phase 2: Hypothesis-Driven Testing

Using the system model, the agents generated and executed test hypotheses across the full interaction surface: navigation flows, media playback, phone pairing sequences, climate control interactions, and settings configurations. Each hypothesis was systematically validated across vehicle trims and connectivity states.

Phase 3: Cross-Language Validation

The agents ran the same exploration and validation sequences across all 38 supported languages — detecting truncated labels, layout overflows, missing translations, and language-specific rendering issues that manual testers had consistently missed.

The Results

• 90%+ reduction in test cycle time — from weeks of manual testing and script maintenance to days of autonomous validation
• Full language coverage — all 38 variants tested systematically for the first time, with defects identified in 12 language packs that had previously passed spot checks
• Zero script maintenance — agents adapted to UI changes across OTA updates without any manual intervention
• Complete auditability — every test action, finding, and system state was logged with full traceability, meeting the OEM's documentation requirements for safety-adjacent systems

What Changed

The OEM integrated Filuta's autonomous agents into their continuous integration pipeline. Every build is now validated automatically, with results available within hours rather than weeks. The QA team shifted from manual test execution to defect analysis and validation strategy — higher-value work that leverages their domain expertise.

Testing went from being our release bottleneck to being our competitive advantage. We ship faster, with more confidence, and our team focuses on what actually matters.

— Head of Software Quality, European Automotive OEM