Intelligent automation (IA) platforms have demonstrated the ability to make predictable, structured, and rule-based human and system work patterns more efficient. Now the challenge turns to agentic processes that require AI for effective deployment. Platforms must add five capabilities: model option and constraint management; content and format processing; the ability to create AI agents; agentic orchestration; and governance, data, and IP protection. While some features of existing IA platforms will become less essential over time, a deterministic process engine will remain indispensable for controlling end-to-end processes for the next three years.