Kyndryl has introduced a new service designed to help organizations prepare their IT environments for the use of autonomous AI agents. With the launch of Agentic Service Management, the company is focusing on automating and managing IT processes in which AI independently performs tasks within complex environments.
According to Kyndryl, many organizations face limitations because their IT structures are still designed for manual processes. This creates a gap between what AI technology is technically capable of and what is actually possible within existing systems. Although companies are increasingly investing in AI, it remains difficult to translate those investments into concrete results.
The new service combines assessments, guidelines, and implementation models to help organizations transition to more autonomous IT processes. It identifies gaps in service management, AI governance, security, and operational processes. Based on this analysis, a phased roadmap is developed to help organizations adapt their IT environment for the use of AI agents.
Kyndryl states that many IT environments are still designed for employees who manually handle tickets and operate systems. In a context where autonomous agents perform tasks across hybrid and multi-cloud environments, the company believes this leads to friction.
Control and governance remain crucial
In an explanatory note, Kyndryl’s strategy lead indicates that organizations must first adapt their operational models before AI can be effectively deployed at scale. This means, among other things, that clear control mechanisms and measurable adoption phases are necessary, while people remain ultimately responsible for governance and risks.
According to Kyndryl, the approach is suitable for a variety of IT environments, including hybrid infrastructures and multi-cloud setups. Kyndryl also cites specific workloads, such as databases and ERP systems, as areas where automation with AI agents can be applied to accelerate modernization.
In addition to implementing agent-driven workflows, Kyndryl emphasizes control and security. In conjunction with the service, the company offers a complementary solution that helps organizations centrally manage and monitor AI agents. This focuses on a centralized overview of active agents to prevent uncontrolled use of AI, as well as on validation prior to deployment. Agents are tested for security, resilience, and compliance with policy rules before going into production.
Control mechanisms remain active even during use. For example, boundaries are set within which agents are allowed to operate, and anomalous behavior can be detected. According to Kyndryl, it is essential that security and governance are part of the development and deployment of AI systems from the start, rather than being added afterward. For example, by incorporating threat modeling and security testing into development processes and applying runtime protection, incidents can be contained more quickly.
Kyndryl is also applying this new approach internally to its own services. Through its proprietary platform, AI functionalities are deployed to support operational processes and assist employees with monitoring and decision-making within business-critical systems.