Reinforcing Real-world Service Agents: Balancing Utility and Cost in Task-oriented Dialogue

TL;DR

This paper introduces InteractCS-RL, a reinforcement learning framework for task-oriented dialogue agents that balances user satisfaction and operational costs through a multi-granularity approach and cost-aware optimization.

Contribution

It presents a novel multi-granularity reinforcement learning framework with cost-aware policy optimization for real-world service agents, addressing complex strategic trade-offs.

Findings

Significantly outperforms baselines in real business scenarios

Demonstrates robustness across diverse domains

Effectively balances user reward and cost constraints

Abstract

The rapid evolution of Large Language Models (LLMs) has accelerated the transition from conversational chatbots to general agents. However, effectively balancing empathetic communication with budget-aware decision-making remains an open challenge. Since existing methods fail to capture these complex strategic trade-offs, we propose InteractCS-RL, a framework that reframes task-oriented dialogue as a multi-granularity reinforcement learning process. Specifically, we first establish a User-centric Interaction Framework to provide a high-fidelity training gym, enabling agents to dynamically explore diverse strategies with persona-driven users. Then, we introduce Cost-aware Multi-turn Policy Optimization (CMPO) with a hybrid advantage estimation strategy. By integrating generative process credits and employing a PID-Lagrangian cost controller, CMPO effectively guides the policy to explore Pareto boundary between user reward and global cost constraints. Extensive experiments on customized real business scenarios demonstrate that InteractCS-RL significantly outperform other baselines across three evaluation dimensions. Further evaluation on tool-agent-user interaction benchmarks verify InteractCS-RL robustness across diverse domains.