OutSystems is introducing a new approach to AI development designed to address a growing problem within organizations: namely, the lack of coherence and control in the use of AI agents. Under the banner of Agentic Systems Engineering, the vendor aims to bring structure to a landscape that is becoming increasingly fragmented.
The timing of the announcement is no coincidence. Companies are actively experimenting with generative AI and tools that produce code autonomously. However, in practice, these applications prove difficult to integrate into existing IT environments. In many cases, isolated initiatives emerge that are hard to scale and have little alignment with existing architectures.
According to OutSystems, a major bottleneck lies in the lack of context. AI agents often operate based on limited information. They do generate functional code, but without taking into account broader dependencies, business logic, or compliance requirements. This increases the risk of fragmentation and makes management more complex.
The proposed approach therefore revolves around explicitly capturing that context. A central role is reserved for the so-called Enterprise Context Graph. This captures relationships between applications, data, processes, and systems. The idea is that AI agents no longer operate in isolation, but within a shared and continuously updated overview of the IT environment.
In doing so, OutSystems addresses a problem that extends beyond the mere deployment of AI. Many organizations are grappling with a growing gap between new applications and existing systems. Legacy environments, integration challenges, and fragmented data make it difficult to actually bring innovations into production.
Role of developers is shifting
The new version of the Mentor development assistant aligns with this. It is designed to support developers in working with AI within this broader context, including through a more dialogue-driven approach to development. The developer’s role is shifting further toward guiding and monitoring AI output, rather than writing all the code themselves.
It is notable that OutSystems takes the use of external AI tools into account. Instead of opting for a completely closed platform, the approach is based on an environment where different agents operate side by side. The condition, however, is that they adhere to the same underlying structure and rules.
This approach underscores a broader trend in the market, where organizations are seeking ways to combine flexibility with manageability. The rapid rise of AI tools makes it easier to experiment. At the same time, it increases the need to maintain control over architecture, data, and governance.
OutSystems plans to launch an early access program in the second quarter of 2026, allowing customers to test the new capabilities.