Document Management

Document Management

Tool Mapping

Document Intake Workflow

Standard Intake Procedure

For every new document:

1. Upload the document via upload_document
2. Review auto-classification — Paperless-ngx will attempt auto-tagging and correspondent matching
3. Verify and correct metadata:
   • Title: Clear, descriptive, includes date if relevant (e.g., "2026-02 Electricity Bill - Tata Power")
   • Date: The document's date (issue date, statement date), not the upload date
   • Correspondent: The organization or person who issued the document
   • Document type: Select from the taxonomy below
   • Tags: Apply all relevant tags from the taxonomy below
4. Confirm filing — Verify the document appears in expected searches

Title Conventions

Use consistent title formatting for searchability:

Tagging Taxonomy

Primary Category Tags

Apply exactly one primary category tag per document:

Secondary Tags (Apply as Many as Relevant)

Tag Hygiene Rules

• Every document must have exactly one primary category tag
• Apply tax-relevant liberally — it is easier to remove than to find missing documents at tax time
• Review needs-renewal tagged documents monthly
• When a policy expires, replace active-policy with expired and remove needs-renewal

Correspondent Management

Correspondent Naming Conventions

• Use the official organization name, not abbreviations: "HDFC Bank" not "HDFC"
• For individuals (doctors, lawyers): "Dr. Firstname Lastname" or "Adv. Firstname Lastname"
• For government: "Income Tax Department", "RTO Hyderabad", etc.

When to Create New Correspondents

• Only create a new correspondent when an existing one does not match
• Before creating, search existing correspondents for partial matches
• Merge duplicates when found (same entity, different name variants)

Document Types

Search Strategies

Finding Documents

Search Tips

1. Use full-text search for content within documents (OCR text)
2. Combine tag filters with date ranges for precise results
3. When searching for a document you know exists but cannot find, try:
   • Search by correspondent instead of title
   • Broaden the date range
   • Search for distinctive terms from within the document content
   • Check if the document was tagged with an unexpected primary category

Retention Policies

How Long to Keep Documents

Retention Review Procedure

Quarterly:

1. Filter documents by expired tag
2. Check if retention period has passed
3. For documents past retention: verify no ongoing need, then archive or delete
4. For utility bills > 1 year old without tax-relevant tag: safe to delete

Integration with Other Skills

Tax Preparation (Budgeting Skill)

At tax time:

1. Search for all documents tagged tax-relevant within the financial year
2. Group by deduction section (80C, 80D, 24, etc.)
3. Verify completeness: each claimed deduction has supporting documentation
4. Cross-reference against actual-mcp transaction data for amount verification
5. Export or share relevant documents

Medical Records (Family)

When visiting a doctor or during medical events:

1. Search for prior medical records by correspondent (hospital/doctor) or tag medical
2. Pull recent lab reports and prescriptions for reference
3. After the visit: upload any new documents and tag appropriately
4. For children: always apply the child-[name] tag

Insurance Claims

When filing a claim:

1. Locate the active policy document: filter by insurance + active-policy + correspondent
2. Gather supporting documents (medical reports for health, FIR for vehicle, etc.)
3. Upload claim submission documents and tag as claim
4. Track claim status by adding notes to the claim document

Common Pitfalls

1. Skipping metadata at upload — Always complete title, date, correspondent, type, and tags immediately. Documents without metadata become unfindable.
2. Inconsistent titles — Follow the title conventions above. A search for "electricity bill" should find all electricity bills.
3. Missing tax-relevant tags — When in doubt, tag it. Removing an unnecessary tag is cheaper than missing a deduction.
4. Correspondent proliferation — Search before creating. "HDFC Bank", "HDFC", and "HDFC Bank Ltd" should be one correspondent.
5. Ignoring document dates — The document date is the date on the document (statement date, issue date), not when you uploaded it. Paperless uses this for date-range filtering.
6. Not reviewing needs-renewal — Check monthly. An expired insurance policy with no renewal is a risk, not just a filing issue.