GALENA: A Governance-Aware LLM Enterprise Navigation Architecture for Autonomous Multi-Agent Workflow Automation with Compliance Enforcement

Paper ID: EIJTEM_2026_13_3_130-146

Author's Name: Narasimha Rao Boinapalli

Volume: 13

Issue: 3

Year: 2026

Page No: 130-146

Abstract:

Enterprise workflow automation increasingly demands intelligent, autonomous systems capable of executing multi-step business processes while maintaining regulatory compliance across heterogeneous data environments. Yet existing LLM-based agentic frameworks fail to enforce governance constraints as first-class architectural invariants, leaving compound action sequences vulnerable to cross-cutting regulatory violations that local, per-step checks cannot detect. This paper presents GALENA (Governance-Aware LLM Enterprise Navigation Architecture), a multi-agent orchestration framework that embeds a formal constraint set Ω into a composite performance scoring function evaluated prior to every agent action deployment, enabling governance-aware rollback and revision loops that provably satisfy GDPR, HIPAA, and domain-specific compliance requirements. The framework introduces three key innovations: (1) an integrated governance gate G(W_i) ≤ τ that enforces policy compliance at every workflow transition rather than at pipeline output; (2) a context-aware agent selection function f_θ(C_q, U_r, H_t) that routes enterprise queries to the optimal specialist agent based on query context, user role, and interaction history; and (3) a drift-resilient lifecycle management mechanism that continuously monitors temporal drift D(W_i, t) and triggers re-evaluation when policy or data distribution changes are detected. Experimental evaluation across three enterprise task categories (Regulated Document Processing, Financial Workflow Automation, and IT Service Management) demonstrates that GALENA achieves 97.0% task completion accuracy, outperforming the strongest baseline (DataGovAgent) by 18.3% (p

Keywords: large language models, enterprise automation, governance-aware agents, multi-agent orchestration, compliance enforcement, autonomous decision-making, LLM pipelines, policy constraint satisfaction

References:

[1] S. Yao, J. Zhao, D. Yu, N. Shafran, K. Narasimhan, Y. Cao, and A. Ibarz, "ReAct: Synergizing Reasoning and Acting in Language Models," in *Proc. 11th Int. Conf. Learning Representations (ICLR)*, Kigali, Rwanda, 2023. [Online]. Available: https://openreview.net/forum?id=WE_vluYUL-X
[2] Q. Wu, G. Bansal, J. Zhang, Y. Wu, S. Zhang, E. Zhu, B. Li, L. Jiang, X. Zhang, and C. Wang, "AutoGen: Enabling Next-Generation LLM Applications via Multi-Agent Conversation," in *Proc. ICLR 2024 Workshop on LLM Agents*, Vienna, Austria, 2024. [Online]. Available: https://arxiv.org/abs/2308.08155
[3] H. Chase, "LangChain: Building applications with large language models," GitHub, 2022. [Online]. Available: https://github.com/langchain-ai/langchain
[4] O. Khattab, A. Singhvi, P. Maheshwari, Z. Zhang, K. Santhanam, S. Vardhamanan, S. Haq, A. Sharma, T. T. Joshi, H. Moazam, H. Miller, M. Zaharia, and C. Potts, "DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines," in *Proc. 12th Int. Conf. Learning Representations (ICLR)*, Vienna, Austria, 2024. [Online]. Available: https://arxiv.org/abs/2310.03714
[5] N. Shinn, F. Cassano, A. Gopinath, K. Narasimhan, and S. Yao, "Reflexion: Language Agents with Verbal Reinforcement Learning," in *Advances in Neural Information Processing Systems (NeurIPS)*, vol. 36, New Orleans, LA, 2023, pp. 8634–8652.
[6] Z. Liu, Z. Han, G. Yan, H. Liang, B. Zeng, X. Chen, Y. Song, and W. Zhang, "DataGovBench: Benchmarking LLM Agents for Real-World Data Governance Workflows," arXiv preprint arXiv:2512.04416, Dec. 2025.
[7] S. Mehta, "Beyond Accuracy: A Multi-Dimensional Framework for Evaluating Enterprise Agentic AI Systems," arXiv preprint arXiv:2511.14136, Nov. 2025.
[8] A. Yehudai, L. Eden, A. Li, G. Uziel, Y. Zhao, R. Bar-Haim, A. Cohan, and M. Shmueli-Scheuer, "A Survey on Evaluation of LLM-based Agents," in *Findings of the Association for Computational Linguistics (ACL)*, 2025. [Online]. Available: https://arxiv.org/abs/2503.16416
[9] Y. Yang, H. Chai, Y. Song, S. Qi, M. Wen, N. Li, J. Liao, H. Hu, J. Lin, G. Chang, W. Liu, Y. Wen, Y. Yu, and W. Zhang, "A Survey of AI Agent Protocols," arXiv preprint arXiv:2504.16736, Apr. 2025.
[10] B. Zhang, M. Takeuchi, R. Kawahara, S. Asthana, M. M. Hossain, G. Ren, K. Soule, and Y. Zhu, "Enterprise Benchmarks for Large Language Model Evaluation," arXiv preprint arXiv:2410.12857, Oct. 2024.
[11] S. S. Sakhinana, V. S. Vaikunth, and V. Runkana, "Knowledge Graph Modeling-Driven Large Language Model Operating System (LLM OS) for Task Automation in Process Engineering Problem-Solving," in *Proc. AAAI Symposium Series*, vol. 4, no. 1, 2024. doi: 10.1609/aaaiss.v4i1.31796
[12] P. Bechard, O. M. Ayala, E. Chen, J. Skelton, S. Davasam, S. Sunkara, V. Yadav, and S. Rajeswar, "Terminal Agents Suffice for Enterprise Automation," arXiv preprint arXiv:2604.00073, Mar. 2026.
[13] M. A. Ferrag, N. Tihanyi, and M. Debbah, "From LLM Reasoning to Autonomous AI Agents: A Comprehensive Review," arXiv preprint arXiv:2504.19678, Apr. 2025.
[14] A. Bucaioni, M. Weyssow, J. He, Y. Lyu, and D. Lo, "A Functional Software Reference Architecture for LLM-Integrated Systems," in *Proc. 22nd IEEE Int. Conf. Software Architecture (ICSA)*, New and Emerging Ideas Track, 2025.
[15] T. Dettmers, A. Pagnoni, A. Holtzman, and L. Zettlemoyer, "QLoRA: Efficient Finetuning of Quantized LLMs," in *Advances in Neural Information Processing Systems (NeurIPS)*, vol. 36, New Orleans, LA, 2023, pp. 10088–10115.
[16] J. Wei, X. Wang, D. Schuurmans, M. Bosma, B. Ichter, F. Xia, E. Chi, Q. Le, and D. Zhou, "Chain-of-Thought Prompting Elicits Reasoning in Large Language Models," in *Advances in Neural Information Processing Systems (NeurIPS)*, vol. 35, New Orleans, LA, 2022, pp. 24824–24837.
[17] Z. Xi, W. Chen, X. Guo, W. He, Y. Ding, B. Hong, M. Zhang, J. Wang, S. Jin, E. Zhou, R. Zheng, X. Fan, X. Wang, L. Li, K. Chen, Z. Yang, B. Shi, F. Zhu, Y. Liu, W. Lai, R. Zhong, D. Li, M. Zhu, C. Li, Q. He, J. Zhang, X. Qin, Z. Wu, Y. Liu, Y. Huang, J. Fu, H. Liu, J. Gui, H. Yang, and T. Gui, "The Rise and Potential of Large Language Model Based Agents: A Survey," arXiv preprint arXiv:2309.07864, Sep. 2023.
[18] G. Sridhara, A. H. Mazumdar, and C. H. Krishna, "Rethinking Legal Compliance Automation: Opportunities with Large Language Models," in *Proc. 32nd IEEE Int. Requirements Engineering Conf. (RE)*, Reykjavik, Iceland, 2024, pp. 41–52. doi: 10.1109/RE59067.2024.00015
[19] A. J. Wang, M. Hu, H. Liu, and T. Lan, "LLM Deployment in Regulated Enterprise AI Systems: A Privacy-Preserving and Compliant Architectural Approach," in *Proc. IEEE Int. Conf. AI and Systems Engineering (ICASE)*, 2025. doi: 10.1109/ICASE.2025.11296166
[20] X. Liu, G. Hu, X. Core, R. E. Welsch, V. Djanatliev, J. Paredis, M. Amin, T. Kraft, and J. K. Allen, "A Reference Architecture for Deploying Large Language Model Applications in Industrial Environments," in *Proc. IEEE Int. Conf. Industrial Informatics (INDIN)*, 2024. doi: 10.1109/INDIN.2024.10814877
[21] T. L. Griffiths, B. M. Lake, A. Ullman, and J. B. Tenenbaum, "Doing More With Less: Meta-Reasoning and Compositional Thinking in Humans and Machines," in *Proc. AAAI Conf. Artificial Intelligence*, vol. 37, Washington DC, 2023, pp. 15683–15689.
[22] M. Chen, J. Tworek, H. Jun, Q. Yuan, H. P. de Oliveira Pinto, J. Kaplan, H. Edwards, Y. Burda, N. Joseph, G. Brockman, A. Ray, R. Puri, G. Krueger, M. Petrov, H. Khlaaf, G. Sastry, P. Mishkin, B. Chan, S. Gray, N. Ryder, M. Pavlov, A. Power, L. Kaiser, M. Bavarian, C. Winter, P. Tillet, F. P. Such, D. Cummings, M. Plappert, F. Chantzis, E. Barnes, A. Herbert-Voss, W. Hebgen Guss, A. Nichol, A. Paino, N. Tezak, J. Tang, I. Babuschkin, S. Balaji, S. Jain, W. Saunders, C. Hesse, A. N. Carr, J. Leike, J. Packard, D. Kirov, E. Slama, S. Eagle, D. Dennison, D. Ziegler, T. Barnes, A. Webber, G. Irving, P. Gao, H. Hernandez, S. Berner, B. McGrew, W. Zaremba, P. Knight, M. Ryan, D. Mane, S. McCandlish, A. Radford, D. Amodei, and W. Zaremba, "Evaluating Large Language Models Trained on Code," arXiv preprint arXiv:2107.03374, Jul. 2021.

View PDF