OIL & GAS REFINING
Although the Oil Refinery held both the asset records and the documents, there was no reliable way to know which documents belonged to a given asset, which assets appeared in a P&ID, which manual applied to a specific equipment record, or which surrounding assets a failure might affect. To gather the information for a single maintenance activity, technicians and planners had to search across SAP, document repositories, shared folders, drawings and manuals.
The Oil Refinery initially considered handing the document-to-asset matching to an external service provider. That approach meant a large team manually inspecting documents, reading off equipment tag numbers, and associating each document with the correct record. For the full scope, the estimate was roughly:
The method was slow, expensive, hard to scale, dependent on manual interpretation, and difficult to maintain as new assets and documents were added. The real problem was not document management. It was the absence of a connected digital relationship between the Oil Refinery’s assets, engineering documents, maintenance history, procedures, spare parts and process dependencies.
Rive designed a Document-to-Tag Matching pipeline that automates the creation of these relationships. Several specialized AI agents work in parallel to process technical documents, identify equipment tags, link them to SAP equipment records, extract asset knowledge, and organize everything within an industrial knowledge graph.
With the graph in place, technicians and planners use Rive as an intelligent maintenance workspace. When a pressure safety valve develops a fault, a technician opens the asset in Rive and, from one interface, reviews the equipment record and specifications, checks previous SAP work orders and when the asset was last maintained, opens the relevant P&ID to understand its process context and connected equipment, retrieves the correct troubleshooting guide and maintenance procedure, and identifies the required spare parts.
The technician can then ask the agent to prepare a work order. The draft can include the equipment reference, the identified problem, recommended activities, relevant procedures, required spare parts, safety and isolation considerations, supporting document references and information from previous work orders. The technician reviews and approves it, and the completed work order is created in SAP following the Oil Refinery’s existing process. Rive does not replace SAP - it provides the intelligence layer that connects documents, asset knowledge and maintenance history before synchronizing the approved transaction back into the system of record.
Rive completed the initial implementation for the Oil Refinery’s first plant. The delivered scope covered a substantial asset base within the first plant and a large engineering and maintenance document repository, including automated tag extraction, matching against SAP equipment records, asset-to-document and asset-to-spare-part relationships, extraction of specifications, procedures and troubleshooting content, linking of historical SAP work orders to their assets, and extraction of asset relationships from P&IDs - all consolidated into an industrial asset knowledge graph with AI-assisted work-order preparation that syncs back into SAP.
The Document-to-Tag Matching pipeline processed a large document repository in a matter of hours.
Targeted human review, focused only on low-confidence matches and exceptions, was completed within a matter of weeks.
The first plant represents a significant portion of the Oil Refinery’s equipment estate and of its engineering and maintenance document repository.
The implementation demonstrated that the same pipeline and knowledge-graph architecture can be extended across the Oil Refinery’s remaining plants.
More case studies
Email us
For general inquiries and information requests.
Book A Demo
Arrange a call to explore Rive’s capabilities and discover how it can meet your operational needs.
