训练语音代理迁移至未见受监管催收流程
原标题：Following the Flow the Law Requires: Training Voice Agents to Transfer to Unseen Regulated Procedures in Debt Collection
AI 摘要
该研究针对受监管场景下的语音代理，提出用机器可读规范 FlowSpec 生成策略指令、工具门控、监督信号和测试，使一个 9B 参数模型在 297 个随机化程序生成的 2,506 段对话上微调后，能迁移到未见过的催收等受监管流程。在泰国外呼催收场景中，文档化检查通过率从 67–71% 提升至 93–100%，场景通过率从 15/43 提升至 40/43，同时审计发现共享规范存在 11 个缺陷。
正文节选
Following the Flow the Law Requires: Training Voice Agents to Transfer to Unseen Regulated Procedures in Debt Collection Abstract Voice agents built on large language models (LLMs) now handle customer calls, but in regulated applications completing the task is not enough: the agent must follow a conversational flow set by law. In debt collection, it must confirm it is speaking to the debtor before mentioning the debt, identify itself and the creditor, state the amount, and avoid prohibited langu