OutSystems, a leading AI development platform, has published the global report State of AI Development 2026. It shows a clear shift where organizations have moved from experimenting with AI to actual use, and AI is increasingly being deployed in mission-critical processes.
Almost all surveyed organizations (96%) are already using AI agents in some form, and 97% are exploring the possibilities of deploying agents more broadly within the organization. The results further show that in the Netherlands, 58%* of organizations are now developing software with the support of genAI, narrowly surpassing traditional code (56%) as the most used method for application development.
At the same time, governance lags behind the rapid adoption. For instance, 94% of organizations indicate that the growth of AI applications leads to more complexity, technical debt, and security risks. Only a small portion of organizations employs a centralized approach to managing agentic AI. In practice, this means that AI agents are often scaled in fragmented environments.
Agentic AI is a next step in the development of AI: systems that autonomously execute workflows, make decisions, and adapt in real-time. According to Gartner, by the end of 2026, 40% of enterprise applications will utilize AI agents for specific tasks. This underscores how quickly this technology is becoming part of enterprise software. In the OutSystems research, which surveyed 1,900 IT decision-makers worldwide, 49% indicate that their organization has agentic AI capabilities at an advanced or expert level.
The Netherlands follows the rise of AI
The degree of adoption varies by region. Organizations in the Netherlands, Australia, Brazil, Germany, the United Kingdom, and the United States show an average maturity level, while France is still in an earlier phase. Particularly in the financial and technology sectors, agentic AI has already been widely adopted in production.
In the Netherlands, it is notable that 58%* of the surveyed organizations are now developing software with the support of genAI, narrowly surpassing traditional code (56%) as the most used method for software development. Legacy systems are seen as the main barrier to successful implementation of agentic projects in our country at 37%, closely followed (36%) by challenging integrations with existing systems. Furthermore, it appears that Dutch decision-makers, compared to other countries, least often cite a lack of skills within their own organization (19%) as a reason for agentic projects not getting off the ground.
The impact of agentic AI is most visible within IT and software development, where the speed at which results are achieved is easily measurable. For instance, 31% of global respondents indicate that AI is now an integral part of how they develop software, and another 42% have integrated AI into specific phases of the software development cycle. As agents prove their value in software development, organizations are shifting to a human-on-the-loop model, where systems operate largely autonomously while humans supervise and can intervene as needed.
"The transition from experimenting with AI to measurable business results is no longer a vision of the future, but the reality of today. The findings from the State of AI Development report show that software development and AI development are increasingly converging," says Woodson Martin, CEO of OutSystems. "As organizations move towards a 'system of agents model', the challenge lies not only in adoption but especially in creating a stable architecture that can manage these complex, intelligent systems and deliver real productivity gains."
"Our approach to deploying OutSystems for an agentic solution was to start with a small, clearly defined project that we knew we could bring to production and that would truly impact the organization," says Scott Finkle, VP of Technology at McConkey Auction Group. "The main goal was to gain experience in developing AI projects for the future. OutSystems and Agent Workbench will pay off significantly as we further develop our AI implementation."
More control over agentic AI
Despite the momentum, fragmentation in architecture remains a significant challenge. For example, 38% of organizations combine self-developed and off-the-shelf agents, leading to AI environments that are difficult to standardize and secure. Only 12% have established a central platform to manage this growth. Most organizations are still experimenting with governance approaches that vary by team and region.
To help organizations gain more control over this development, OutSystems recently introduced Agentic Systems Engineering: an approach to AI development that supports organizations in building, managing, and evolving controlled agentic systems for organizations. Download the State of AI Development 2026 report and discover how organizations are making the transition from experimenting with AI to actual use.