Over the past two years, we’ve explored ADKAR through the lens of AI adoption—and agentic patterns through the lens of change management. Here’s the synthesis: a framework that combines both.

ADKAR, Adapted for AI

Awareness: Understand that the bottleneck has shifted from implementation to judgment. Understand the risks, dependencies, and upside. Build ongoing awareness—it deepens with experience.

Desire: Want to participate in the new way. Desire grows from early wins, reduced friction, psychological safety, and narrative alignment (queue-to-craft). It’s organizational as well as individual—product and tech need aligned desire.

Knowledge: Learn what matters when agents execute—prompt design, evaluation, orchestration, failure modes. Deprioritize syntax and procedure; prioritize judgment and workflow.

Ability: Develop the skills to direct, validate, and integrate agent output. For ICs: judgment, prompts, testing, glue work. For PMs: scope judgment, spec prompts, evaluation, orchestration. Ability includes the bridge from prototype to production.

Reinforcement: Sustain the change. Bake agents into workflows. Design for defaults and feedback loops. Reinforce human roles—strategy, judgment, coherence, learning—so they remain central. Plan for non-linear, iterative change.

Agentic Patterns, Layered In

Augmentation → Automation → Orchestration: Map your initiatives to the taxonomy. Know where you are and what each step requires in terms of ADKAR.

Delegation, Supervision, Handoff: Design the human-agent boundaries. When do humans delegate? When do they supervise? How do handoffs work? These patterns shape Knowledge and Ability.

Multi-Agent Orchestration: When you scale to multiple agents, invest in boundaries, handoff contracts, human checkpoints, and observability. Ability and Reinforcement both depend on it.

Conflict Resolution: Plan for agent-agent and agent-human conflict. Define who decides what. Design escalation paths. Reinforcement includes sustaining those protocols.

The Integrated View

ADKAR guides people through change. Agentic patterns guide system design. They’re orthogonal but complementary. Use ADKAR to prepare and sustain the org. Use agentic patterns to design the workflows people will adopt. The framework that works is the one that does both—explicitly, and in coordination.

This isn’t the last word. The tools will evolve; the patterns will refine. But the combination of change management (ADKAR) and system design (agentic patterns) gives teams a structure to navigate what’s ahead. Use it, adapt it, and share what you learn.