# A Governance-Aware Large Language Model Orchestrated Agentic Digital Twin for Transmission System Operator Control Room Decision Support

arXiv: https://arxiv.org/abs/2609.22476 · HTML: https://mlsanity.com/paper/2609.22476/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agentic AI for wireless networks, one of 16 directions in LLM agents and tool use.
- Builds on 15 of its 36 references that are in the map.
- First posted on arXiv 18 Sep 2026.

## Placed in

- LLM agents and tool use / [Agentic AI for wireless networks](https://mlsanity.com/agents/agentic-ai-for-wireless-networks/)

## Builds on (most cited first, 15 of 15)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks](https://mlsanity.com/paper/2211.12588/) (2022)
- [PAL: Program-aided Language Models](https://mlsanity.com/paper/2211.10435/) (2022)
- [Defeating Prompt Injections by Design](https://mlsanity.com/paper/2503.18813/) (2025)
- [AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents](https://mlsanity.com/paper/2503.18666/) (2025)
- [Language Models Do Hard Arithmetic Tasks Easily and Hardly Do Easy Arithmetic Tasks](https://mlsanity.com/paper/2406.02356/) (2024)
- [GridMind: LLMs-Powered Agents for Power System Analysis and Operations](https://mlsanity.com/paper/2509.02494/) (2025)
- [X-GridAgent: An LLM-Powered Agentic AI System for Assisting Power Grid Analysis](https://mlsanity.com/paper/2512.20789/) (2025)
- [Towards an AI assistant for power grid operators](https://mlsanity.com/paper/2012.02026/) (2020)
- [PowerChain: A Verifiable Agentic AI System for Automating Distribution Grid Analyses](https://mlsanity.com/paper/2508.17094/) (2025)
- [Facilitating AI and System Operator Synergy: Active Learning-Enhanced Digital Twin Architecture for Day-Ahead Load Forecasting](https://mlsanity.com/paper/2409.00368/) (2024)
- [NRT-Bench: Benchmarking Multi-Turn Red-Teaming of LLM Operator Agents in Safety-Critical Control Rooms](https://mlsanity.com/paper/2606.20408/) (2026)
- [A Conversational Agentic Interface to Physics-Based Household Digital Twins for Residential Energy Decision Support](https://mlsanity.com/paper/2606.31744/) (2026)

## Abstract (arXiv)

Transmission system operators face rising complexity from renewable integration, reduced inertia, and tighter security margins. Large language models offer natural-language decision support, but their hallucinations, uncontrolled tool use, and weak traceability conflict with control room requirements. This paper presents a governance-aware agentic digital twin for transmission grid control rooms. The large language model only selects and parameterizes whitelisted analysis tools, and every proposed action passes through a governance layer that the model cannot bypass. The layer enforces four rules on every run. Only whitelisted tools execute. No run exceeds its step budget. No action with side effects executes without explicit operator approval. Every number in an answer is rendered by the layer from backend results with its unit, variable, and time. The rules are checked on a persistent audit trail for every run of a released 118-task benchmark, which covers analytics, simulation, multi-step workflows, and twelve families of adversarial inputs on a digital twin of the Greek transmission network. Across 590 runs of the primary model, tool selection reaches 96.5% and task success 93.7%, and all four rules hold without exception. In a separate three-repetition study across four large language models, 1416 runs in total, the rules again hold on every run, with an approximate 95% lower bound of 99.8%. Removing the layer makes the same model execute all 45 approval-requiring runs without authorization and leaves only 39.2% of its answers with backend-supported numbers. Enforcement costs 12 to 16 milliseconds per request.