As we approach the end of 2026, organizations are facing a daunting challenge: proving the value of their artificial intelligence (AI) investments. With trillions of dollars spent on AI deployments over the past year, companies must now demonstrate not only that their AI systems are secure but also that they’re generating tangible returns on investment (ROIs). Failure to do so could have severe consequences for organizations that have come to rely heavily on AI.
The problem is more pressing than ever, with reports of autonomous AI agents acting without human oversight continuing to emerge. This has sparked a heated debate about the need for greater control and governance over AI systems, which are increasingly complex and difficult to manage. As Melinda Marks, practice director of cybersecurity at Omdia, notes, “Organizations must take proactive steps to best manage and protect data, set policies and guardrails for secure access and usage, and secure the software supply chain as it complexity increases.”
The issue is not just about security; it’s also about accountability. With AI budgets expected to increase by 10% or more next year, companies will be under pressure to demonstrate their investment’s worth. Omdia predicts that success in 2027 will be measured by ROIs and productivity gains, rather than technical superiority alone. This means that organizations must develop metrics to track the financial returns of their AI initiatives and be prepared to justify their spending.
Chief information security officers (CISOs) are particularly well-positioned to address these challenges, but they face significant hurdles. Gartner analysts warn that more than half of organizations lack a defined approach to limit AI agent access, leaving them vulnerable to potential risks. To mitigate this risk, CISOs must prioritize governance and infrastructure protection, ensuring that their AI systems are designed with security in mind from the outset.
Gartner’s advice is clear: governing autonomous multiagent systems requires a fundamental shift in how organizations approach AI adoption. By implementing robust policies and guardrails, companies can ensure that their AI agents operate within defined limits and reduce the blast radius of potential security incidents. As Gartner notes, “An AI system does not need to achieve artificial general intelligence capabilities to create significant cyber risk. It simply needs the ability to impact enterprise operations.”
In light of these challenges, organizations must take a proactive approach to AI governance and security. This means developing clear metrics for measuring ROI, investing in robust infrastructure protection, and implementing policies that govern AI agent access. By doing so, companies can not only ensure their AI systems are secure but also demonstrate the tangible value they bring to the organization.
Ultimately, the era of AI accountability is upon us, and organizations must be prepared to answer the question: what’s the return on investment from our AI investments?
Source: Dark Reading — 2026-10-02