At the Databricks AI Summit in San Francisco, the central theme was that AI is moving from a back-office tool to the front lines, with contact centers directly in its path. The conversation focused on agentic AI—systems that can reason, plan, and execute multi-step tasks without human intervention for every decision. This shift enables automating complex interactions like billing disputes and technical troubleshooting. Enterprises are using Databricks’ Unity Catalog and Mosaic AI to build governed, auditable AI pipelines, with emphasis on data governance, lineage tracking, access controls, and model monitoring to ensure responsible deployment. However, a gap remains between controlled demos and live customer interactions at scale, with hallucination, context window limits, and underestimated organizational change management posing challenges. Companies succeeding are moving deliberately, piloting in low-risk channels, measuring obsessively, and building human-AI collaboration models before scaling. For CX leaders, the architectural decisions being made today about how AI agents are trained are the same ones their technology partners are building on.