While companies are rapidly adopting AI agents into their development environments, issues regarding security, governance, and integration with complex existing systems still remain during the actual operation phase. To mitigate these limitations, OutSystems introduced 'Agentic System Engineering,' proposing a direction that goes beyond the use of individual generative tools to manage the entire enterprise system within a controllable structure.

OutSystems announced this concept on April 6. According to the company, the 'Enterprise Context Graph' is at the heart of the new approach. This graph is designed to identify the relationships between applications, data, workflows, and agents in real time, and aims to help AI agents operate based on the necessary context, even within complex systems.

The key lies in ensuring that AI goes beyond simply generating code quickly to enabling development while understanding the structure and rules of the enterprise environment. OutSystems explained that this allows them to maintain structural consistency and support development that meets security and regulatory compliance requirements. The vision is to ensure that developers work within the same context and guardrails, even when using multiple tools.

This structure is implemented across the platform through the next-generation 'Mentor'. Mentor was introduced as a tool that combines interactive application creation with development experience within an integrated development environment (IDE) to support the building and enhancement of complex systems with fewer iterations. The company stated that this environment is not limited to its own platform but can be integrated with various agentic coding tools.

Initial case studies featured clients in the university, IT services, manufacturing, and distribution sectors. Kent State University stated that it is driving operational efficiency improvements by reducing repetitive tasks such as code exploration, documentation, and error diagnosis, while Balantic explained that it is utilizing the technology to establish initial project structures more quickly. SRS Distribution revealed that it reduced the development time for some applications by up to 50% and disclosed a case where an attendance management app was transitioned to the operational phase in approximately two weeks after planning.

This announcement can also be interpreted as an example showing that enterprise AI development is shifting from a competition in single-model performance to a competition in actual operating systems. OutSystems announced plans to launch an early access program for customers in the second quarter of 2026.