Connecting & Managing Documents

Documentation Index

Fetch the complete documentation index at: https://docs.programmeinsights.co.uk/llms.txt

Use this file to discover all available pages before exploring further.

​

Before you start

• You are logged in to Programme Insights (see guide 00).
• You have Reviewer or Admin permissions on the workspace.
• If using SharePoint as a source, the SharePoint site and folder must be accessible to the Programme Insights service principal (configured by your admin).

​

Key concept: connect, don’t upload

Programme Insights follows a connect, don’t upload philosophy. Rather than copying files into the platform, you bind it to the location where your documents already live — typically a SharePoint folder. This means:

• Documents stay in your organisation’s existing document management system.
• Version control and access permissions are managed in SharePoint as usual.
• The platform reads from the source; it does not modify your originals.
• You can re-sync at any time to pick up new or updated files.

Manual upload is available as a fallback for documents not stored in SharePoint.

​

Steps

​

Viewing the documents page

1. Open an assessment and click Documents in the sidebar.

[screenshot: 13-documents.png]

​

Connecting a document source

2. Click Connect Source and choose one of:

   Source type	How it works
   SharePoint folder	Provide the SharePoint site URL and folder path. The platform connects via Microsoft Graph API using your organisation’s service principal. All files in the folder (and subfolders) are discovered.
   Manual upload	Drag and drop files or browse to select. Files are stored in the platform’s own storage. Use this for documents outside SharePoint.

3. Confirm the connection. The platform begins discovering and processing files.

​

Understanding document states

4. Each document passes through a processing pipeline:

   State	Meaning
   Uploaded	File has been received or discovered but processing has not started.
   Processing	Text extraction and indexing are in progress.
   Processed	Document is fully indexed and ready for assessment.
   Error	Processing failed (e.g. corrupt file, unsupported format). Check the error message for details.

​

What happens during processing

5. Behind the scenes, each document goes through three stages:
   • Text extraction — Azure AI Document Intelligence extracts text, tables, and structure from PDFs, Word documents, and other supported formats.
   • Chunking — Extracted text is split into semantically meaningful chunks for retrieval.
   • Embedding — Each chunk is vectorised and stored for similarity search during assessment.

​

Re-syncing documents

6. If documents in your SharePoint folder have been added, updated, or removed, click Re-sync to refresh the connection. The platform will:
   • Discover new files and begin processing them.
   • Re-process files whose content has changed.
   • Mark removed files as unavailable (they will not be deleted from previous assessment results).

​

Reviewing match status

7. Once an assessment has run, the documents page shows additional columns:
   • Match Status — whether the document was matched to any criteria during the assessment.
   • Criteria Coverage — how many criteria reference evidence from this document.

​

What happens next

• Create an assessment that uses these documents: Creating a New Assessment
• Run the assessment and watch it process: Running & Monitoring Assessments

​

Common questions

Q: What file formats are supported? PDF, DOCX, XLSX, PPTX, and plain text. Scanned PDFs are supported via OCR through Azure AI Document Intelligence. Q: How long does document processing take? Typically 1-3 minutes per document depending on size and format. Large PDFs with many pages may take longer. Q: Can I connect multiple SharePoint folders to one assessment? Currently each assessment connects to a single document source. If you need files from multiple locations, consolidate them into one SharePoint folder or use manual upload for the additional files.