cs.AIMar 31, 2026

CausalPulse: An Industrial-Grade Neurosymbolic Multi-Agent Copilot for Causal Diagnostics in Smart Manufacturing

Authors: Chathurangi Shyalika, Utkarshani Jaimini, Cory Henson, Amit Sheth

Organizations: Artificial Intelligence Institute, University of South Carolina · Stride Lab, University of Michigan, Dearborn · Bosch Center for Artificial Intelligence, Pittsburgh

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Modern manufacturing environments demand real-time, trustworthy, and interpretable root-cause insights to sustain productivity and quality. Traditional analytics pipelines often treat anomaly detection, causal inference, and root-cause analysis as isolated stages, limiting scalability and explainability. In this work, we present CausalPulse, an industry-grade multi-agent copilot that automates causal diagnostics in smart manufacturing. It unifies anomaly detection, causal discovery, and reasoning through a neurosymbolic architecture built on standardized agentic protocols. CausalPulse is being deployed in a Robert Bosch manufacturing plant, integrating seamlessly with existing monitoring workflows and supporting real-time operation at production scale. Evaluations on both public (Future Factories) and proprietary (Planar Sensor Element) datasets show high reliability, achieving overall success rates of 98.0% and 98.73%. Per-criterion success rates reached 98.75% for planning and tool use, 97.3% for self-reflection, and 99.2% for collaboration. Runtime experiments report end-to-end latency of 50-60s per diagnostic workflow with near-linear scalability (R^2=0.97), confirming real-time readiness. Comparison with existing industrial copilots highlights distinct advantages in modularity, extensibility, and deployment maturity. These results demonstrate how CausalPulse's modular, human-in-the-loop design enables reliable, interpretable, and production-ready automation for next-generation manufacturing.

Figures & tables

Figure 1: High-level overview of CausalPulse system architecture and components.
Figure 2: CausalPulse four-layer framework: user-facing, agent, utility, and data, connected through standardized protocols (MCP, A2A, and LangGraph).
Figure 3: Illustration of domain-specific rules for a grinding process. Control parameters (rectangles) may cause observation parameters (circles) at the same or later grinding stage (Rules R1, R2, R5), observation parameters may only cause later observations (Rules R3, R4), and edges violating type or temporal order are explicitly disallowed (dashed red arrows).
Figure 4: Agentic task orchestration and workflow execution in CausalPulse: (a) Query interpretation and planning by the LangGraph-based workflow engine, (b-I) Agent card registration and central server coordination, (b-II) Agent interactions and data flow during execution, and (b-III) Postprocess chaining with the Recommender Agent.
Figure 5: Root cause analysis results: (a) ranked root-cause paths ordered by causal significance for the anomaly in E_8 in the PSE dataset (refer to ”Datasets Descriptions” Section), and (b) visual representation of the corresponding causal graph, highlighting identified root causes and propagation paths. The legend on the right lists variable names and their hierarchical levels, where levels represent the sequential order of grinding steps in the PSE use case.
Figure 6: Canonical agent pipeline in CausalPulse. Predefined workflows correspond to subpaths of this pipeline with optional reuse of cached intermediate results.
IDWorkflow path (agents in order)
W1Preprocessing →\rightarrow Recommender
W2Preprocessing →\rightarrow Background Info →\rightarrow Recommender
W3Background Info →\rightarrow Recommender
W4Preprocessing ∗ →\rightarrow Anomaly Detection
 →\rightarrow Recommender
W5Preprocessing ∗ →\rightarrow Background Info →\rightarrow
Table 1: Predefined workflows in CausalPulse as paths through the canonical agent pipeline in Figure 6 . Steps marked ∗ may reuse cached results if available.
DatasetStageCriteriaNo.of QueriesExpected (steps/agents)Predicted (steps/agents)Successful attemptsFailed attemptsCriterion success (%)Aggregate criterion success (%)
PSEPreprocessingPlanning2033200100.098.73
Tool use2022200100.098.73
Self-reflection203219195.098.73
Multi-agent collab.203319195.098.73
Anomaly detectionPlanning5033500100.098.73
Tool use5033500100.098.73
Table 2: Criterion success and aggregated criterion success per stage across datasets, workflow stages, and agentic criteria
Figure 7: Aggregated success rate per agentic criterion for PSE and FF datasets.
Anom.Ground-truth causesPRORCA top predictionsHits@kPrec.Rec.F1
F4{ SetAngle_3 , GrindDepth_3 , GrindDepth_4 }{ SetAngle_4 , GrindDepth_4 }hits@1 = 0, hits@2 = 10.500.330.40
E12{ GrindDepth_E_9 , GrindDepth_E_12 }{ SetAngle_3 , GrindDepth_E_12 , GrindDepth_E_9 , SetAngle_4 , SetAngle_2 }hits@1 = 0, hits@2 = 1, hits@3 = 1, hits@4 = 1, hits@5 = 10.401.000.57
E3{ SetAngle_3 , GrindDepth_3 }{ GrindDepth_3 , SetAngle_3 , SetAngle_4 , GrindDepth_4 }hits@1 = 1, hits@2 = 1, hits@3 = 1, hits@4 = 10.501.000.67
Table 3: Representative PRORCA root-cause attribution outcomes on the PSE dataset, illustrating (top) a partial match where only one of three ground-truth causes appears in the top-2 positions, (middle) a high-recall but noisy case where all ground-truth causes are recovered along with several false positives, and (bottom) a strong match where both ground-truth causes are ranked in the top-2 positions.
 UCSDMATMCDAssetOpsSmartPilotCausalPulse
1. System typeSingle agentMulti-agent discoveryMulti-agent simulationMulti-agent copilotIndustry-grade multi-agent copilot
2. End-to-end     
causal diagnostics     
pipeline✗✗✗✗✓
3. Dynamic     
multi-agent     
Table 4: Summary comparison of CausalPulse with representative industrial copilots.
Figure 8: End-to-end deployment pipeline of CausalPulse
More in this topic

Smart Manufacturing

See all 10 papers

Explore similar work

CardsList
Jul 24, 2026

Agentic Root Cause Analysis through Evidence-Grounded Reasoning

Diagnosing the root cause of anomalies is essential for safe industrial operation. Despite extensive sensor instrumentation, formulating hypotheses and gathering evidence remains a manual process, creating a major operational bottleneck. While existing data-driven approaches aim to automate this, two critical limitations restrict their deployment: their operate as black boxes unable to justify their diagnosis, and they require scarce labeled examples of faulty operation. To address this gap, we introduce AgentRCA, a zero-shot agentic framework for evidence-grounded root cause analysis. Rather than learning fault-specific mappings, AgentRCA performs inference-time reasoning by combining a data-driven digital twin (modeling normal system dynamics) with a tool-augmented large language model. The agent iteratively gathers statistical evidence, evaluates competing hypotheses, and identifies the physical fault that best explains the observed behavior. Evaluated on a real-world multiphase-flow facility and a large-scale chemical plant, AgentRCA achieves diagnostic performance competitive with fully supervised baselines without relying on fault-specific training. Crucially, it produces transparent reasoning traces that explicitly link observed symptoms to their underlying physical causes. These results establish autonomous hypothesis-driven reasoning as a practical foundation for scalable industrial root cause analysis.

Amaury Wei, Olga Fink

EPFL - IMOS Laboratory CH-1015 Lausanne Switzerland

Industrial Anomaly DetectionAgentic ReasoningRoot Cause Analysis
AskPDF
May 26, 2026

ORCA: An End-to-End Interactive Copilot for Optimized Root Cause Analysis

Causal analysis is a crucial task in many domains, including manufacturing, social science, and medicine. However, despite recent progress, the conceptual and methodological complexity of causal methods makes them largely inaccessible to domain experts. This gap prevents experts from leveraging these advances and hinders researchers who lack access to real-world data for validation. To bridge this divide, we introduce ORCA, a copilot for end-to-end causal analysis. ORCA orchestrates agents to understand the user's goals and guide them through the most appropriate causal analysis workflow, from fully automatic to highly user-guided execution. It features causal discovery, causal effect estimation, explainability and Root-Cause-Analysis (RCA). ORCA evaluates and compares performance, generates key metrics and diagrams, and generates insights through structured reports. We highlight its effectiveness across several real-world use-cases.

Phi Nguyen Xuan, Nicholas Tagliapietra, Lavdim Halilaj +2

Bosch Global Software Technologies Company Limited, Vietnam · Robert Bosch GmbH, Germany · Computer Science Department, TU Darmstadt, Germany +2

AI-Assisted Decision MakingCausal DiscoveryTool-Using AgentsRoot Cause Analysis
AskPDF
Oct 1, 2026

LLM-Driven Multi-Agent Control for Skill-Based Smart Manufacturing

Factories are shifting toward smaller lot sizes with high product customization, requiring frequent re-programming of flexible and reconfigurable automation systems. LLM-based agents can be deployed in two complementary roles: Offline, they generate deterministic production sequences, reducing programming effort; online, they operate live machines and handle unforeseen runtime faults that static programs cannot anticipate. We propose a solution in which each factory module is paired with a dedicated LLM-based agent and an MCP tool server that exposes the module's skills via OPC UA method calls, with agents coordinating over MQTT and grounded by real-time updates of the factory state. We compare three agent architectures (orchestrator, peer-to-peer, and monolithic) across nine production challenges of increasing complexity in a simulation of a physical six-module hexagonal factory, including silent hardware fault detection. The monolithic and peer-to-peer architectures both achieve the highest mean solve rate (93%), while the orchestrator uniquely resolves a silent conveyor-belt fault in all ten runs by autonomously rerouting plates around the blocked segment. All architectures exhibit emergent fault-diagnosis behavior without any explicit failure-handling logic, establishing standardized MCP tooling, MQTT-based inter-agent communication, and real-time state injection as a viable and reproducible foundation for LLM-programmed smart manufacturing.

Kay Köhle, Darko Anicic, Thomas A. Runkler +1

Technical University of Munich, Munich, Germany · Siemens AG, Foundational Technologies, Munich/Garching, Germany · Siemens AG, Digital Industries, Fürth, Germany

ManufacturingMulti-Agent LLM SystemsLLM Agent OrchestrationMulti-Agent ControlFault DetectionSmart Manufacturing
AskPDF