Towards Enforcing Company Policy Adherence in Agentic Workflows
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909825104871424 |
|---|---|
| author | Zwerdling, Naama Boaz, David Rabinovich, Ella Uziel, Guy Amid, David Anaby-Tavor, Ateret |
| author_facet | Zwerdling, Naama Boaz, David Rabinovich, Ella Uziel, Guy Amid, David Anaby-Tavor, Ateret |
| contents | Large Language Model (LLM) agents hold promise for a flexible and scalable alternative to traditional business process automation, but struggle to reliably follow complex company policies. In this study we introduce a deterministic, transparent, and modular framework for enforcing business policy adherence in agentic workflows. Our method operates in two phases: (1) an offline buildtime stage that compiles policy documents into verifiable guard code associated with tool use, and (2) a runtime integration where these guards ensure compliance before each agent action. We demonstrate our approach on the challenging $τ$-bench Airlines domain, showing encouraging preliminary results in policy enforcement, and further outline key challenges for real-world deployments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_16459 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Enforcing Company Policy Adherence in Agentic Workflows Zwerdling, Naama Boaz, David Rabinovich, Ella Uziel, Guy Amid, David Anaby-Tavor, Ateret Computation and Language Large Language Model (LLM) agents hold promise for a flexible and scalable alternative to traditional business process automation, but struggle to reliably follow complex company policies. In this study we introduce a deterministic, transparent, and modular framework for enforcing business policy adherence in agentic workflows. Our method operates in two phases: (1) an offline buildtime stage that compiles policy documents into verifiable guard code associated with tool use, and (2) a runtime integration where these guards ensure compliance before each agent action. We demonstrate our approach on the challenging $τ$-bench Airlines domain, showing encouraging preliminary results in policy enforcement, and further outline key challenges for real-world deployments. |
| title | Towards Enforcing Company Policy Adherence in Agentic Workflows |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2507.16459 |