OutSystems, a low-code AI development platform, released its 2026 State of AI Development report on April 12, surveying 1,900 IT leaders globally. The headline finding: 96% of enterprises are already using AI agents in some capacity, and 97% are exploring organization-wide agentic AI strategies. But 94% of those same organizations say agent sprawl is increasing complexity, technical debt, and security risk.
The Numbers
The survey, conducted between December 2025 and January 2026, paints a picture of an industry that has moved well past experimentation. According to PRNewswire, 49% of respondents describe their agentic AI capabilities as “advanced” or “expert.” Financial services and technology organizations report the highest levels of production deployment.
Within IT and software development, 31% say AI is integral to their development practices, and 42% have embedded AI into specific phases of the software development lifecycle. More than half (52%) have shifted to a human-on-the-loop model, allowing agents to operate with reduced direct oversight while maintaining supervisory control.
Governance Remains Fragmented
The concerning counterpoint: only a small fraction of enterprises have established centralized governance for their agent deployments. Most are running agents across fragmented environments with inconsistent oversight, according to the report. Thirty-eight percent of organizations mix custom-built and pre-built agents, creating AI stacks that are difficult to standardize and secure. Only 12% have implemented a centralized platform to manage sprawl.
This lines up with a Gartner prediction that 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% in 2025. If 94% of current adopters already report governance anxiety at today’s scale, that number is heading in one direction as deployment accelerates.
Regional Variation
APAC is outpacing other regions. India reports some of the highest levels of advanced and expert agentic AI capability, according to SD Times. Australia and Japan are moving from pilot to production. Organizations in Brazil, Germany, the Netherlands, the UK, and the US report intermediate progress. France remains earlier in its adoption curve.
Context and Caveats
OutSystems is a vendor with a direct commercial interest in this space. The company recently launched OutSystems Agentic Systems Engineering, a platform for building governed agentic systems. Vendor-conducted surveys tend to skew toward respondents already engaged with the vendor’s product category, which could inflate adoption figures.
That said, the 94% governance concern figure is consistent with similar findings from other sources. Kore.ai’s separate State of AI 2026 report found 89% of enterprises plan to increase AI investment, while KPMG found only 11% have deployed agents at enterprise scale. The gap between deployment velocity and governance maturity is showing up across multiple independent surveys.
The Governance Supply Chain
The timing of this report matters. In the past week alone, Microsoft open-sourced an Agent Governance Toolkit covering all 10 OWASP agentic AI risks. Trent AI emerged from stealth with $13M to build runtime security for autonomous agents. AWS launched an Agent Registry for enterprise governance. The supply side is responding to exactly the demand signal OutSystems quantifies here: enterprises have agents in production, but no centralized way to track, secure, or govern them.
The question is whether governance tooling can scale as fast as agent deployment. At 96% adoption and 12% centralized governance, the gap is already wide. Every additional agent deployed without oversight compounds the technical debt and security surface that 94% of respondents are already worried about.