ARdena: Scenario-driven control of real-time LLM agents
For developers of conversational AI agents, this provides a lightweight method to adapt agent behavior dynamically, though the approach is incremental as it relies on structured prompting rather than novel control mechanisms.
The paper introduces a scenario-driven prompting framework for controlling LLM agents in real-time interactive environments, enabling runtime behavior modification without model fine-tuning. The ARDena agent demonstrates that scenario definitions alone produce substantially different interaction behaviors while maintaining stable real-time operation.
Large language models (LLMs) have enabled increasingly capable conversational agents, but reliably controlling their behavior in real-time interactive environments remains a significant challenge. Existing approaches often rely on model fine-tuning or alignment procedures that are difficult to adapt to changing interaction requirements. This paper introduces layered scenario-driven LLM control, a framework that enables runtime behavior control through structured prompting. By combining persistent context with scenario-specific constraints, the approach allows agent behavior to be modified during interaction without changing the underlying model. The framework is implemented in ARDena, a real-time multimodal embodied agent that integrates speech interaction, visual perception, tool use, and avatar-based response generation. The proposed approach is evaluated with respect to control effectiveness, response latency, and operational stability. The results demonstrate that scenario definitions alone can produce substantially different interaction behaviors while maintaining stable real-time operation, highlighting the effectiveness of scenario-driven prompting for controlling LLM agents.