1. Automated Tagging During Ingest

• Every file (PDF, Excel, email, Word, etc.) is parsed at point of ingest.
• The system runs a multi-layered tagging engine:
  • LLM-based content extraction
  • Regex + rule-based patterns
  • Embedding similarity match
  • Custom heuristics based on document type

2. Tag Categories Applied by ViewOps

• Context Tags

  Tags describing operational setting or workflow:

  • Production Batch
  • Change Request
  • NPD Launch
  • Non-Conformance
  • Planning Issue

• Topic Tags

  Tags describing the subject matter or business object:

  • Product Code: 40536
  • Delivery Date: 22/10/2025
  • Yield Percentage
  • Case Weight
  • Customer: Tesco

• Privacy Tags

  Tags marking sensitive or protected content:

  • Contains PII
  • Contains Pricing
  • Confidential
  • Employee Name
  • Internal Only

• Permission Tags

  Tags controlling visibility and access:

  • Viewable by: Ops
  • Restricted to: Quality Team
  • External Share Blocked
  • Audit Required

• Department Tags

  Used for routing, filtering, and dashboards:

  • Dept: Technical
  • Dept: Retail
  • Dept: Finance
  • Dept: Planning

3. What the User Sees (UI/UX)

• Tag sidebar showing all tags with their values, grouped by category.
• Each tag links to the exact highlighted text in the document or email.
• Users can:
  • ✅ Confirm a tag (marks as correct)
  • ❌ Remove a tag (sends feedback to model)
  • (Later: ✏️ Edit tags and values manually)