OutSystems: AI Advances, but Governance Struggles to Keep Up

The Rise of AI in Enterprises and Its Challenges

A survey conducted by OutSystems, titled The State of AI Development 2026, reveals that artificial intelligence (AI) is now in the early production phase in many companies, primarily within IT departments. The study, which gathered responses from 1,879 IT leaders, highlights a concerning gap between the rapid adoption of AI and organizations' ability to govern it effectively. The authors of the report emphasize the need to establish safeguards for AI systems and to integrate these new technologies into existing platforms.

OutSystems indicates that 97% of respondents are exploring agentic strategies, with 49% considering themselves to have "advanced" or "expert" capabilities. Nearly half of those surveyed claim that more than half of their agentic AI projects have moved from the pilot phase to production. Indian companies stand out particularly, with 50% reporting that their AI projects are successful at a rate of 51% to 75%.

Companies are contemplating the optimal location for deploying AI agents and the necessary controls. While cost reduction and efficiency gains are common expectations, only 22% of companies have found their deployments to be effective in these areas. The most notable gains come from equipping developers with generative AI tools, which significantly enhance productivity.

Leading Sectors and Integration Challenges

The financial services and technology sectors are at the forefront of the transition from pilot to production, with implementations in key business functions. These sectors demonstrate a clear understanding of automation and return on investment. The survey reveals that AI-assisted development is common in nine of the ten countries surveyed, complementing traditional development methods.

OutSystems finds that 48% of respondents consider integration with legacy systems crucial for expanding agentic AI, while 38% cite these systems as the main cause of project stagnation. Integration difficulties and the fragmentation of legacy systems are perceived as major obstacles by more than 40% of participants.

Organizations contemplating large data cleanup programs, often recommended by AI vendors to explain deployment failures, may need to rethink their approach. According to the report, agents can be built to operate effectively even in complex data environments, provided that governance and integration are strengthened alongside AI implementation.

Trust and Governance of AI

Trust in agentic AI is improving, with 73% of respondents expressing high or moderate confidence in the autonomy of agents, an increase of 10% from the previous year. Trust in code or workflows generated by third-party AI tools is also on the rise, reaching 67%, up from just 40% the previous year. However, only 36% of companies have centralized AI governance, and 64% do not have such a structure. 41% rely on ad hoc rules, and two-thirds of respondents find it technically challenging to integrate human checkpoints.

Many organizations appear to be deploying looser oversight models, although it is unclear whether this results from greater trust in the models or if business functions are under pressure to deploy AI regardless of security or reliability concerns. If the trend toward relaxing oversight continues, the report's authors note that the adoption of agentic AI could advance more quickly than the accountability methods that many deem important.

Companies wishing to deploy agents in regulated or mission-critical environments should consider orchestration and auditability as integral parts of the product. When compliance controls take into account a company's operations, audit trails in the form of log files and defined responsibilities are seen as essential elements of any agentic AI deployment. The report indicates that 94% of leaders are concerned about the "expansion of AI," which is not defined but could be interpreted as a lack of a centralized management platform overseeing all AI deployments within the company. 39% are very or extremely concerned about this issue, and only 12% currently use a centralized platform to keep this expansion under control.