Kyndryl has launched
Agentic Service Management, a structured approach designed to help enterprises move from traditional IT service operations to AI-driven workflows. The offering combines a maturity model, assessments, and implementation blueprints to guide organizations as they introduce agentic AI into their environments while maintaining control over security, governance, and service outcomes.
The gap between AI capability and IT reality
AI capabilities are moving fast, but most enterprise environments are not built to support them. IT systems still rely heavily on human-led processes - tickets, escalations, and manual coordination. As organizations begin to introduce autonomous agents, that mismatch becomes a real operational problem.
Kyndryl’s own data highlights this gap. Many organizations are investing in AI but struggling to see returns because their operating models have not evolved. The issue is less about access to AI and more about whether the environment can support it at scale.
Why digital trust becomes critical
This is where digital trust starts to matter.
Kris Lovejoy, Global Head of Strategy at Kyndryl, explained to ChannelE2E that the focus needs to shift from individual tools to how the system operates as a whole.
“In the era of agentic AI, digital trust is about an organization’s ability to govern, control, and operate autonomous systems reliably, securely, and at scale, rather than simply measuring the performance of individual agents in isolation.”
Autonomous agents can plan, reason, and act on their own. That changes the risk profile. Traditional controls were not designed for systems that make decisions independently.
“As enterprises transition from human-driven to autonomous decision-making, the primary challenge is establishing an operating model that ensures these systems remain transparent and secure.”
Closing the readiness gap
Kyndryl frames the challenge as a readiness issue. According to its
2025 Readiness Report, 46% of organizations are not seeing meaningful returns from AI because their environments are not built for agentic requirements.
That gap shows up in areas like identity, authorization, and accountability. Without those in place, autonomous systems are difficult to control.
“Organizations must treat AI as an industrialized asset.”
Lovejoy outlines a two-step path to address this. First, modernize the underlying infrastructure so it can support agentic workloads. Second, build an operational foundation that defines how those systems are governed.
“Closing this gap requires updating IT infrastructure to support agentic workloads, and establishing a control plane through assessments and blueprints to ensure that the environment can reliably support autonomous agents.”
What this means for MSPs and service delivery
For managed service providers, this shift opens up a new layer of work. Enterprises are not just looking for tools; they need help managing how AI operates over time.
“The most immediate opportunity lies in ongoing AI lifecycle management.”
Customers are focused on compliance, risk, and control. That creates demand for services that go beyond deployment. Partners can step in to assess readiness, implement governance frameworks, and manage AI systems across hybrid and multi-cloud environments.
“Policy as code and lifecycle management platforms like Kyndryl Bridge allow partners to deploy agents that are inherently trusted and transparent through automated policy enforcement.”
From experimentation to operations
Kyndryl is also applying these ideas internally through its
Bridge platform, where agentic capabilities are being integrated into service delivery. The goal is to reduce manual coordination while keeping human oversight in place.
Kyndryl is positioning this approach as an extension of its experience running large-scale systems.
“Our approach is rooted in decades of managing mission-critical systems without disrupting business as usual. By combining that with our Agentic AI Framework, we are enabling organizations to move from experimental AI to industrial-scale operations with governance and control built in.”
Where this leaves the market
This is about an operational shift. Enterprises are no longer blocked by access to AI. The challenge is making it work reliably inside real environments. As agentic capabilities become more common, the focus will move toward governance, control, and measurable outcomes. Vendors can provide the tools, but organizations still need a way to run them. Kyndryl’s launch points to where that conversation is heading next - away from pilots and toward the question of how AI actually operates at scale.
