AI agents in live operations require new standards and management (www.techradar.com)

AI agents are rapidly integrating into live operations, prompting a critical need for new management standards within enterprises. A recent Kyndryl Report highlights a sharp disparity: while 87% of business leaders foresee AI reshaping career paths, only 29% believe their teams can effectively utilize AI, with many still stuck in an experimentation phase. For AI agents to succeed, organizations must establish clear governance structures, decision rights, and workflows to manage their complexity. The shift from merely deploying technology to orchestrating it effectively is pivotal; poorly defined operational frameworks could lead to inefficiencies and increased risks. As agents scale, they bring challenges across four key strains: data, integration, operational, and governance. Successful implementation requires robust controls, including decision boundaries and real-time monitoring to ensure alignment with business objectives. Organizations must also evolve their testing approaches to account for the dynamic nature of these systems, moving beyond static evaluations to encompass behavior testing. Emphasizing bounded autonomy—balancing agent independence with robust oversight—is crucial for unlocking the full potential of AI agents in enterprise environments. By fostering early collaboration among stakeholders and re-engineering governance and operational models, organizations can prepare to harness autonomous AI effectively and safely.