Datasets / AI Centre of Excellence
Policy and Procedure Corpus
Approved policies and procedures, chunked for retrieval.
158 downloads
Updated weekly. Last update 2 days ago. Next expected in 5 days.
About this dataset
The Bank's approved policies, standard operating procedures, product manuals, circulars and a curated set of public BNM policy documents, chunked with section headings, version and effective date for retrieval. Superseded versions are kept but flagged. The grounding corpus for the policy copilot and the regulatory clause retriever.
Usage rights
Open to all staff for search and GenAI grounding. Answers must cite the source document and version. Documents marked Confidential by their owner are excluded from this corpus and held separately.
Lineage
Pulled weekly from the policy management system and the document repository, converted to text with layout-aware parsing, chunked by section and embedded with a Vertex AI text embedding model. Each chunk keeps its document id, version and effective date. Dataplex checks for documents past their review date and orphaned chunks.
Ownership and sensitivity
Who approves access
No approval needed. Self-service: subscribe and use, logged for chargeback.
Built on by Lumen Risk Labs delivery team.
In Safe to try
Open to all staff. Aggregates, policy text, public series and synthetic data.
Self-service: subscribe and use, logged for chargeback.
2 models trained on this dataset
Details
- Updated
- 2026-08-01
- Cadence
- Weekly
- Format
- PDF + JSONL chunks
- Licence
- Group Reuse
- Latest version
- 2026.31