CRM hygiene and auto-enrichment

Requirements describe capabilities the pattern needs in your environment, not the vendors you must buy. Any system that fills a requirement satisfies it — that’s what makes the catalog portable across the long tail of SMB tooling.

1. crm_record_store

   REQUIREDREAD + WRITErequest

   The CRM whose records are being cleaned. The pattern reads and writes here with care: it never deletes, only flags or merges with audit trail.

   DATA SHAPE

   Contact, company, and deal records with standard fields plus custom fields. Pattern needs write access for updates and a way to log the change source.
   COMMONLY FILLED BY

   • CRM system with API write access
   • internal sales database the team uses as a CRM
   • spreadsheet acting as a CRM (degrades pattern value significantly; consider this a precursor)
2. enrichment_data_source

   REQUIREDREADrequest

   Where missing fields get filled from. Could be one provider or several stacked.

   DATA SHAPE

   Per-contact and per-company data: job titles, company size, industry, funding, recent news. Source attribution and confidence per field.
   COMMONLY FILLED BY

   • subscription data enrichment service
   • free LinkedIn or company-data lookup APIs
   • internal historical activity logs
   • import of an enrichment file from a list provider
3. change_audit_log

   REQUIREDWRITEevent

   Every change the pattern makes is logged here. Critical for trust: a CRM that quietly changes data is worse than one that doesn't try.

   DATA SHAPE

   Per-change entries: record affected, field changed, old value, new value, source of new value, confidence, timestamp.
   COMMONLY FILLED BY

   • audit log inside the CRM
   • separate log table the pattern writes to
   • structured change feed posted to a team channel
4. low_confidence_review_queue

   REQUIREDWRITEbatch

   Changes the pattern isn't confident enough to make automatically go here for a human to approve or reject.

   DATA SHAPE

   Proposed changes with confidence scores, evidence, and approve/reject/edit-and-approve actions for a human.
   COMMONLY FILLED BY

   • review queue inside the CRM
   • dedicated review UI built for the pattern
   • weekly digest emailed to RevOps with batched changes
5. merge_authority

   REQUIREDREAD + WRITEbatch

   Who or what authorizes record merges when duplicates are found. Merges are destructive in effect even if the audit trail exists.

   DATA SHAPE

   Merge proposal with duplicate records, similarity score, suggested primary record, and human or rule-based decision.
   COMMONLY FILLED BY

   • RevOps reviewer with a merge queue
   • rule-based auto-merge above a confidence threshold (only for high-confidence cases)
   • team lead reviewing weekly batches