From Business Events to Auditable Decisions: Ontology-Governed Graph Simulation for Enterprise AI
career value
168K/year
🤖 AI Summary
Current large language model agents often generate responses directly within open knowledge spaces, lacking explicit modeling of how business events dynamically reshape decision contexts, which undermines auditability and reliability. To address this, this work proposes the LOM-action framework, introducing an event-driven enterprise ontology simulation mechanism that deterministically evolves knowledge graphs within an isolated sandbox to produce a scenario-valid simulated graph \( G_{\text{sim}} \). Decisions are then grounded in \( G_{\text{sim}} \), establishing a closed-loop pipeline of “event → simulation → decision.” The framework employs a dual-mode architecture—skill mode and reasoning mode—to ensure full traceability throughout the decision process. Experimental results demonstrate that its toolchain achieves an F1 score of 98.74%, substantially outperforming Doubao-1.8 and DeepSeek-V3.2 by 24–36%, thereby validating the critical role of ontology governance and event-driven simulation in enabling trustworthy enterprise AI decision-making.
Application Category
📝 Abstract
Existing LLM-based agent systems share a common architectural failure: they answer from the unrestricted knowledge space without first simulating how active business scenarios reshape that space for the event at hand -- producing decisions that are fluent but ungrounded and carrying no audit trail. We present LOM-action, which equips enterprise AI with \emph{event-driven ontology simulation}: business events trigger scenario conditions encoded in the enterprise ontology~(EO), which drive deterministic graph mutations in an isolated sandbox, evolving a working copy of the subgraph into the scenario-valid simulation graph $G_{\text{sim}}$; all decisions are derived exclusively from this evolved graph. The core pipeline is \emph{event $\to$ simulation $\to$ decision}, realized through a dual-mode architecture -- \emph{skill mode} and \emph{reasoning mode}. Every decision produces a fully traceable audit log. LOM-action achieves 93.82% accuracy and 98.74% tool-chain F1 against frontier baselines Doubao-1.8 and DeepSeek-V3.2, which reach only 24--36% F1 despite 80% accuracy -- exposing the \emph{illusive accuracy} phenomenon. The four-fold F1 advantage confirms that ontology-governed, event-driven simulation, not model scale, is the architectural prerequisite for trustworthy enterprise decision intelligence.
Problem
Research questions and friction points this paper is trying to address.
LLM-based agent systems
ungrounded decisions
audit trail
enterprise decision intelligence
illusive accuracy
Innovation
Methods, ideas, or system contributions that make the work stand out.
ontology-governed simulation
event-driven decision making
auditable AI
graph mutation
enterprise decision intelligence
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
