Identity management lags behind AI: IDC research

According to a recent study by IDC, published on September 21, 2026, the number of machine identities within organizations is growing rapidly. This leads to significantly increased risks in identity management, of which 90% of surveyed IT decision-makers are aware. The study surveyed 539 senior IT decision-makers and C-level executives, and the figures are alarming: on average, there are 109 machine identities per employee in corporate networks, consisting of service accounts, bots, and API keys.

The traditional methods of identity management, where employees typically received a single account with fixed permissions, are no longer sufficient. The explosive growth of machine identities makes it nearly impossible to keep track of these manually. IDC points out that organizations are not well-prepared for the complexity this situation brings. The need for improvement is urgent; 58.7% of respondents indicate that they need to fundamentally revise their approach.

Although many organizations see the need for action, they lag in implementing effective measures. Only 26.7% of respondents utilize dynamic role-based access control, which is essential in this new reality of AI agents and automated processes. Furthermore, nearly 57.7% lack clear procedures for incidents, while no less than 98.4% of respondents consider the collaboration between IT and security to be suboptimal. All of this creates a vicious circle of uncertainty and risk.

To address these issues, IDC proposes evolving towards an integrated operational process called Resilience Operations (ResOps). This process should consist of a cohesive approach for prevention, detection, response, and recovery, ensuring collaboration between IT, security, and the business. IDC predicts that within three to five years, ResOps will become a standard business function, but until then, the challenge for organizations remains identity management, which urgently needs improvement.

Read the full article from AI Insider.