TRACER-7B: multi-turn user simulator aligning simulated behavior with real interaction trajectories

KEY POINTS

1. The paper (arXiv:2609.28690v1) introduces TRACER, a multi-turn user simulator trained by supervised fine-tuning on real dialogues followed by multi-turn reinforcement learning that uses hierarchical outcome- and trajectory-level rewards and deviation-aware advantage modulation.
2. TRACER-7B outperforms the strongest baseline by 11.4 conversion F1, achieves the lowest group-level conversion-rate error and semantic trajectory distance on real customer-service cohorts, generalizes to out-of-distribution scenarios, and passes human Turing tests near chance; the authors also present the Dynamic Marketing Benchmark, which jointly measures persuasion effectiveness and response quality and finds that higher response quality does not necessarily imply higher conversion rates.
3. More behaviorally consistent user simulators improve training, evaluation, and deployment of interactive AI and reveal that response quality and persuasion outcomes can diverge.