AI Center of Excellence

The rollout team disbands. The champions return to their day jobs. The Slack channel goes quiet. Six months later, adoption has plateaued and new hires are not trained. This is the sustainability gap — the period after initial rollout enthusiasm fades and before AI usage becomes genuinely embedded in organizational culture. The AI Center of Excellence (CoE) bridges this gap.

The CoE is a small team (2-4 people, often part-time) that owns: the prompt library, the training curriculum, the governance framework, the adoption metrics, and the support channel. They do not use Claude for the organization — they enable the organization to use Claude effectively. Their job is maintenance, not innovation. Update the prompt library when workflows change. Train new hires. Review governance incidents. Report adoption metrics to leadership. The CoE makes adoption self-sustaining instead of dependent on the enthusiasm of the original rollout team.

1. 1. Staff the CoE Assign 2-4 people part-time (20-30% allocation). Choose from pilot graduates who demonstrated both skill and enthusiasm. The CoE needs practitioners, not managers.
2. 2. Define the Charter The CoE owns: prompt library maintenance, new hire training, governance oversight, adoption metrics reporting, and escalation support. The charter is specific — ambiguous charters produce ambiguous results.
3. 3. Report Monthly Monthly report to the executive sponsor: adoption rate by department, time savings metrics, governance incidents (if any), prompt library growth, and support ticket trends. Keep the sponsor engaged with data.