BOM Intelligence: AI-Powered Manufacturing Process Optimisation Solution (Manufacturing)

Ontology-driven BOM process automation: significantly faster generation, superior validation accuracy, POC-verified

Forbes AI Top 50 Enterprise

Fortune China Tech Top 50 Enterprise

IDC MarketScape: Generative AI Data Infrastructure — Leaders Position (Top 1 among non-cloud vendors, 2025)

IDC China Manufacturing Big Data Solutions Market Share — Ranked 4th overall, No.1 among non-cloud vendors (2023)

Cases of Dazu Smart Control and Xingsen Technology selected for IDC China Industrial Leadership Award

Member Enterprise of Hong Kong Software Industry Association

Hong Kong Deepexi Technology Limited

Unit 213A, Building 5W, 5 Science Park East Avenue, Hong Kong Science Park, Shatin, New Territories, Hong Kong

1-on-1 Consultation

67363463

chenying01@deepexi.com

https://https://www.deepexi.com/

Summary

This solution applies Deepology ontology modelling and the Deepexi enterprise LLM to automate and validate BOM (Bill of Materials) data for manufacturers. Three FastAGI intelligent agents are deployed: an Engineering Configuration Table Generation Agent, a Consistency Validation Agent, and a Completeness Validation Agent, automating the full flow from product planning BOM through engineering configuration table to MBOM. In a POC with a leading automotive manufacturer, efficiency was significantly improved across all stages and validation accuracy consistently met business requirements. The solution supports private on-premise deployment to ensure data security and control.

Applicable Industry

Solution Category

Generative AI Robotics/Hardware Enterprise Resource Planning (ERP)Workflow Automation Supply Chain Management Inventory Management Product Lifecycle Management (PLM)Business Intelligence (BI)Artificial Intelligence (AI)Greater Bay Area (GBA)

Highlights

Engineering configuration table generation: Significantly faster than manual processing (which requires 32-40 hours), with AI completing generation rapidly. Validation accuracy meets business standards and business user spot-checks confirmed expected resu

Engineering configuration table completeness validation: approximately 25x faster per sub-module (manual: 180-210 minutes, AI: 7-8 minutes). Validation accuracy is stable and consistently meets business benchmarks.

MBOM completeness validation: approximately 48-50x faster for full-table validation (AI completes in 15-20 minutes). Multi-round testing demonstrated stable accuracy performance meeting business requirements.

Deepology ontology modelling preserves BOM domain knowledge as reusable AI data assets, supporting complex business logic reasoning and extension to core BOM types including EBOM and MBOM.

FastAGI agents support private on-premise deployment to protect sensitive manufacturing data and integrate with existing PLM systems.

Service Packages and Offerings

Details

[Core Capabilities] Built on ontology knowledge graph + enterprise LLM + context engineering, delivering intelligent automation across the full BOM lifecycle. The solution can be optimised for generation efficiency and validation accuracy based on the enterprise's actual engineering complexity. It intelligently identifies logical conflicts between engineering and planning configuration tables (e.g., mutually exclusive options simultaneously selected), detects missing or surplus components row-by-row across MBOM and EBOM, and reduces manual validation errors — helping enterprises improve BOM processing efficiency and quality.

[Technology Architecture] Foundation: FastData Foil AI-Ready Data Fusion Platform (multimodal BOM data parsing and format conversion); Core: Deepology Ontology Modelling (entity-attribute-relation triples, dynamic graph maintenance) + Deepexi Enterprise LLM (high-accuracy, zero-hallucination); Application: FastAGI workflow Agents (Generation Agent + Validation Agent + multi-round PDCA iteration); Deployment: Private on-premise; data stays within the enterprise.

[Application Scenario] AI-driven validation and auto-generation of full-vehicle BOM data (engineering configuration BOM, EBOM, MBOM) for automotive manufacturers; extensible to more vehicle models and BOM types.

[Applicable Government Departments] Architectural Services Department; Civil Aviation Department; Electrical and Mechanical Services Department; Innovation and Technology Bureau

Customer Cases

CASE

A leading automobile manufacturer

A leading new-energy automotive manufacturer conducted a full-vehicle BOM data AI validation POC. After deploying FastAGI agents, efficiency was significantly improved across engineering configuration table generation, MBOM completeness validation, and engineering table completeness validation, with stable accuracy across all stages meeting client business requirements. POC conclusion: AI generation and validation are feasible at business scale, with efficiency and accuracy meeting business requirements, ready to extend to more vehicle models and BOM types.

Search Details