OntoL Ontology Case Study: Contract & Payment Verification for Chemical Manufacturer
A case study of OntoL ontology deployment at a chemical manufacturer, detailing two workflows—project-contract-payment querying and contract attachment verification—highlighting data governance gaps, DeepSeek Harness prototyping, and the birth of OntoL-Data governance tooling.
Project Background
The engagement was for a chemical manufacturing company with a poor on-site data environment: no data governance, no dedicated business data personnel, and ad-hoc system integrations.
First Workflow: Project, Contract, Payment Query
The first line of work used three ontologies—Project, Contract, and Payment—to query payment status across projects. A demo built with DeepSeek Harness demonstrated this workflow, and the client was satisfied with the results.
Second Workflow: Contract Compliance Verification
The second line embedded a button in a chat interface to verify contract issues. It compared formal attachments against the contract content displayed on the page. The ontology definitions included:
Contract and Compliance Risk : queries current structured contract information.
Contract Attachment Retrieval for Comparison : queries contract attachment information.
Additional ontologies for verification content.
Prototyping with DeepSeek Harness
The initial demo for the first workflow was created using DeepSeek Harness, which the author used for analysis and presentation. The client responded positively.
Client Feedback and Second Phase Requirements
After the successful first demo, the client requested a mobile-accessible, login-free page for the second workflow. Because the client lacked data governance, the team built an intermediate library management system, which evolved into the OntoL-Data ontology governance product.
Building OntoL-Data Governance System
Development took about a week. The client's data quality was low, requiring several custom interfaces. The author notes the process was challenging.
UI Design Iterations
Initial UI designs were revised after client feedback. The article includes before-and-after screenshots showing the transformation.
Final Testing and Results
After multiple test rounds, the final verification flow worked as demonstrated: clicking the "Verify Contract Content" button produced the expected comparison results.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
AI Large-Model Wave and Transformation Guide
Focuses on the latest large-model trends, applications, technical architectures, and related information.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.