Databricks reports that AI usage in companies is rapidly evolving from simple generative models to agent-based systems. Agentic AI is increasingly planning and executing complex workflows independently. New usage data shows a 327 percent increase in multi-agent implementations on the Databricks platform between June and October 2025.
This suggests that AI is increasingly becoming part of the operational architecture and replacing pilot projects. The so-called supervisor agent plays a central role, coordinating tasks and distributing them to specialized sub-agents. The trend extends beyond tech companies and requires more powerful infrastructure, governance, and sophisticated workflow design in many industries worldwide.