Governance as the Handrail – March 2026
Incrementally introducing agentic-AI within a controlled framework
According to McKinsey’s CIO’s face a choice in modernising enterprise IT systems for AI adoption – incremental change or full-scale transformation. Incremental change allows companies to deploy AI-agents into the technology stack but at the risk of increasing technical debt. For large organisation’s managing systems built over decades, incremental adoption is the only realistic option.
Companies can start by introducing AI-enabled automation into high-value workflows. Early AI adopters have already benefited from this approach, with each AI agent enabling new value to be extracted from old assets. Each step of the modernisation process feeds directly into the next stage introducing enhanced levels of modernisation efficiency.
Governance as a guardrail. McKinsey warns that governance is needed to ensure the smooth transition to AI-enabled automation. This is to ensure all AI deployments operate within defined operational boundaries. McKinsey proposes the introduction of what they call the “AI-agent mesh” to enable the newly automated distributed systems to be managed and controlled.
This aligns with what we call the Topology Control Plane. This is a control layer built incrementally as AI is implemented. The Topology Control Plane enables AI to be integrated in a controlled and observable manner. The impact of each proposed AI change is carefully tested in the Topology Control Plane synthetic environment before it is introduced into production.
This approach changes the risk profile of incremental AI process modernisation. Instead of increasing technical debt, each AI deployment is tested and validated before system change is introduced. This ensures system integrity is preserved and business continuity risk is contained.
As McKinsey’s highlights, just bolting on AI doesn’t generate real enterprise value. Only by introducing AI in a controlled and continuously managed framework can AI enablement add value and scale at a measured pace