Institutional workflow patterns

This page outlines practical ways departments and central offices are using XML upload to scale CPOS compliance.

Pattern A: Admin-owned data, PI-owned certification (recommended)

1. Admin maintains a structured CPOS inventory (spreadsheet, database, …)
2. Admin exports/upload-prep XML for each PI (one file per individual)
3. PI uploads XML into SciENcv CPOS document
4. PI reviews, completes missing fields, and certifies
5. Admin collects the SciENcv-generated certified PDF for submission packaging

Pattern B: Hybrid drafting (admin seeds, PI edits)

• Admin uploads baseline XML (projects, dates, sources)
• PI adds narrative fields (objectives, overlap) and final effort estimates

Pattern C: System integration

Some institutions generate SciENcv-ready CPOS XML directly from internal research administration systems (e.g., effort reporting, award management, disclosure tooling). If you do this:

• Treat awardnumber as the primary identifier for de-duplication and updates.
• Generate person-month rows deterministically (inclusive year range).
• Build a validation step (see validator script in this repo) before handing files to PIs.

Governance tips

• Maintain a clear cutoff date for when updates are accepted before submission.
• Use an intake form for items that are hard to infer (consulting, in-kind, foreign appointments).
• Keep a copy of the exact XML uploaded, and the final certified PDF that was submitted.

flowchart LR
    accTitle: Admin-owned data and PI-owned certification
    accDescr: Institutional XML workflows can centralize data preparation while keeping review and certification with the PI or named individual.
    A["Structured CPOS inventory"] --> B["Per-person XML export"]
    B --> C["Upload or seed SciENcv CPOS"]
    C --> D["PI reviews and completes entries"]
    D --> E["PI certifies"]
    E --> F["Admin collects certified PDF"]
    F --> G["Archive XML and submitted PDF"]