Contact-file field mapping

Map by meaning and evidence, not header similarity.

Identity fields

• source_row_id: stable identifier added before cleaning.
• profile_url_raw and profile_url_normalized: preserve original and canonical URL.
• full_name_raw: original display value.
• first_name_normalized, last_name_normalized: populate only when parsing is defensible.
• company_name_raw: original company value.
• company_domain_normalized: lowercase domain with protocol and www removed.

Preferred identity order:

1. canonical professional-profile URL;
2. existing verified work email;
3. normalized full name plus company domain;
4. normalized full name plus unambiguous company name.

Rows below level 4 need review rather than blind lookup.

Result fields

Append rather than replace:

• work_email_candidate
• work_email_status
• phone_candidate
• phone_status
• enrichment_source
• enriched_at
• enrichment_job_id
• enrichment_note

If the original field is empty and the result meets the approved quality threshold, copy the candidate into the delivery column. If the original is non-empty and differs, preserve both and set enrichment_note = conflict_review.

Duplicate handling

Assign a duplicate_group_id. Look up one canonical identity per group, but return every source row. Duplicate removal is a separate user decision.