Change Management in the AI Era Part 2: Adaptive Leadership in Practice

Two pressure points are breaking AI transformations: governance built for a slower world, and a trust gap between what leaders say and what people experience. These are not separate problems — they feed each other. Most transformations that lose momentum die here.

• 82% of organizations say AI risks have accelerated the need to modernize governance, yet most still run pre-AI oversight (OneTrust, 2025).
• 42% of companies abandoned most AI initiatives in 2025, up from 17% the year before (S&P Global, 2025).

The Governance Gap

Approval cycles, oversight committees, and annual planning were designed for change that moved in phases. AI doesn’t work that way. It reshapes tasks, roles, and decision rights continuously. Agentic systems now act without human review. Three questions are already pressing:

• Who is accountable when AI influences a hiring or promotion decision?
• Which decisions must stay human, regardless of what the system recommends?
• How fast can the organization catch and fix a problem before it reaches a KPI?

The principle is simple: AI doesn’t decide people’s salaries or careers. Humans do. Operationalizing that at speed is the harder problem. Organizations doing it well have stopped treating governance as compliance and started running it as an operating function:

• Cross-functional from day one. HR, Legal, Risk, and Ops at the table before decisions are made — not after.
• Tiered by risk. Deep review for anything touching people, customers, or regulators. Light-touch for lower-stakes experiments.
• Paced to the work. When workflows change quarterly, annual governance isn’t governance.

The Trust Gap

Trust is slow to build and quick to collapse — and it’s tested every time leadership communication and employee experience diverge. In AI transformations, they diverge constantly. Managers are asked to lead through ambiguity they weren’t trained for. CHROs hear the same thing from their teams: give us something stable enough to commit to.

Builder.ai is the cautionary case. The company raised over $450 million on an AI platform that turned out to be largely human engineers. When that gap surfaced, the company went bankrupt. The dynamic is the same inside organizations. When what leaders say keeps diverging from what people experience, trust doesn’t erode gradually. It collapses.

Trust is rebuilt through consistency — how leaders behave when transformation gets hard, long after the announcement is made. One CHRO in the Council Advisors network sends junior HR staff to talk directly with AI users after major rollouts. Not to check compliance. To listen.

Why the Two Gaps Compound

Slow governance erodes trust. Eroded trust makes governance slower. Decisions drag. Reasoning goes opaque. People fill in their own version of what’s happening — and it’s rarely charitable. Some teams stop waiting and build workarounds, which carry their own risks. The loop doesn’t need a dramatic event to close. It closes through accumulation. This is why well-resourced transformations led by capable people still lose ground.

The Takeaway

This isn’t a process problem. It’s a leadership problem. The answer is visible, specific, and consistent leadership — past the announcement, past the deployment, past the first hard quarter.

Leadership Action Plan