AI interfaces for financial portfolio modeling and optimization are evolving from graphical user interfaces toward conversational, agent-driven systems. At SS&C Algorithmics, we are developing a multi-agent AI framework for end-to-end financial risk management. Users can deploy teams of AI agents to define risk-factor scenarios, propagate their impacts to the instrument and portfolio levels, and run multiple portfolio optimization strategies to evaluate risk-reward trade-offs and design hedges. Multi-agent AI systems and conversational interfaces are not only augmenting and increasingly replacing traditional graphical user interfaces but are also evolving into advanced decision-support systems that financial institutions can use to plan, evaluate, and validate financial decisions. We will explore multi-agent architectures, share practical lessons learned, and discuss open challenges in ensuring the reliability, transparency, and validation of agentic AI systems in financial applications.