As the use of generative AI in embedded software development increases, test automation tools are also expanding beyond supporting simple repetitive tasks to directly connect requirements with verification results. Vectorcast 2026, announced by Vector Korea, is a platform that adds a requirements-based unit test automatic generation feature to align with this trend.

Vector Korea announced on March 30 the launch of 'Vectorcast 2026,' an embedded software test automation platform featuring an AI-powered requirements-based test generator. The company explained that this version aims to enhance test traceability and simplify certification and quality assurance procedures for safety-critical systems.

The key feature is the newly introduced 'Reqs2x'. This feature analyzes software requirements, links them to the functions that implement them, and automatically generates executable unit tests that meet those requirements. Requirements information is designed to be imported from common management tools such as DOORS, Polarion, and CSV files.

Program slicing and large-scale language models are utilized together in the test generation process. This reduces the discrepancy between requirements specifications and actual implementations, and enables clearer verification of which functions have been verified according to which requirements. However, the generated mappings and test cases are not automatically finalized but are subject to user review.

This aligns with the awareness that while the introduction of generative AI increases productivity, it also requires additional control and supervision in safety-critical development environments. Vector stated that it has ensured that AI-generated outputs are subject to the same level of verification and quality control procedures as code written by humans. The company explained that this structure also helps align with industry standards, such as MISRA C:2025, ISO and IEC AI frameworks, and the EU AI Act.

In terms of operation, it supports a BYOM model that allows customers to directly select their preferred AI models. Accordingly, companies can deploy in a form that fits their policies among on-premises, cloud, and hybrid environments, and manage data privacy, compliance, and cost control internally. Vector provides configuration connectors, while the actual AI infrastructure is operated by the customer.

Vectorcast 2026 is designed to integrate with automated quality monitoring pipelines, combining test result analysis and reporting capabilities. Given its use in safety-critical fields such as aerospace, automotive, medical, industrial control, and railway, this update is interpreted as focusing on strengthening test systems that combine automation, traceability, and human review. The Reqs2x feature is available to existing Vectorcast 2026 license holders at no additional cost.