As businesses rush to embrace AI-powered software development, new research from OutSystems reveals a stark challenge: balancing the promise of agentic AI with mounting security, governance, and compliance risks.
The global study of 550 software executives found that 93% of organisations are either building or planning to develop their own custom AI agents—autonomous tools designed to automate processes, analyze data, and drive decision-making. But while early adopters report gains in productivity, scalability, and software quality, the rapid rise of AI agents is introducing a new layer of complexity to enterprise security.
According to the report, 64% of software leaders now view security, governance, and compliance risks as top concerns associated with AI adoption. Many cite challenges with controlling the transparency and reliability of AI-driven decisions, particularly as organisations scale AI agents across workflows and applications. Further, 44% of respondents pointed to the growing problem of AI sprawl, where disconnected AI tools increase technical debt and introduce oversight gaps.
“AI agents have tremendous potential to automate and accelerate digital transformation—but only if organisations apply the same rigor to AI governance as they do to traditional systems,” said Woodson Martin, CEO of OutSystems. “Without careful oversight, AI proliferation can introduce vulnerabilities, compliance failures, and opaque decision-making processes that threaten business integrity.”
While the promise of agentic AI includes hyper-personalised experiences, automated processes, and seamless enterprise-wide data access, its power also demands new approaches to security architecture. AI agents’ ability to autonomously access and act on data raises concerns around data privacy, regulatory compliance, and unauthorized decision-making.
To address these risks, OutSystems recommends that businesses establish clear governance frameworks around AI use, invest in workforce upskilling, and prioritize transparency in AI design and operations.
The report also forecasts a wave of workforce transformation, as nearly 70% of software executives expect AI adoption to create new specialist roles—such as AI governance leads and agent architects. Furthermore, 63% believe significant reskilling of current development teams will be required to manage AI-driven systems securely.
Despite the challenges, most organisations remain committed to AI-driven development, seeing it as essential to maintaining competitive advantage. With customer service leading current deployments, executives are exploring AI agents in areas like product development, supply chain management, and HR—underscoring the need for cross-enterprise security strategies.
“Security teams must be at the table from the beginning,” said Michael Harper, Managing Director at KPMG LLP. “AI’s value isn’t just in automation—it’s in trusted automation. Businesses that build security and governance into their AI strategies from day one will be those that unlock AI’s full potential.”
OutSystems’ study serves as a clear warning: as AI agents become more autonomous and embedded, safeguarding systems, data, and decision-making processes is no longer optional—it’s mission-critical.
