CRM Automation Audit Checklist: 25 Things to Review Before You Scale

Before adding sales staff, review lead routing, duplicate records, lifecycle triggers, and sync failures. Problems in these workflows can affect both daily work and pipeline reporting.

This checklist gives you the specific things to review before a major headcount expansion, a CRM migration, or a RevOps overhaul.

Pipeline and Stage Configuration

1. Every pipeline stage has a written entry criterion. If you cannot write it in one sentence based on a verifiable CRM event, the stage is ambiguous. Ambiguous stages make forecasts harder to interpret.

2. Every pipeline stage has a stale-deal rule. If a deal can sit in any stage indefinitely without triggering a task or an alert, stale deals can inflate the forecast.

3. No stage allows deals to skip multiple levels automatically. If automation can advance a deal from "Prospect" to "Closed Won" in one step, something is wrong with either the trigger logic or the stage structure.

4. Closed Lost reasons are standardized and enforced. If "Closed Lost" has no required reason field, your loss analysis is useless. Enforce this at the workflow level.

5. Deal value fields are required before stage X. Every deal that reaches a proposal stage with no deal value is a forecasting hole. Make the field required before that transition.

Workflow and Trigger Logic

6. Every active workflow has a documented purpose and owner. If you cannot answer "why does this workflow exist and who maintains it," deactivate it. Record the purpose and dependencies before deciding whether to disable a workflow.

7. Enrollment triggers are as specific as possible. "Contact created" is not a valid trigger for a sales notification workflow. It should be "Contact created AND lifecycle stage = MQL AND company size > 100."

8. Every workflow has exit conditions and suppression logic. Check every active workflow: what exits a contact from it? Is there a suppression list that prevents customers from receiving lead nurture?

9. No two workflows enroll the same contact for the same reason simultaneously. Run an enrollment overlap test: take five recent contacts and trace which workflows they are or were enrolled in. Overlapping enrollments that serve the same purpose are a signal.

10. Workflows that send email check send-time settings. A 3am email because a contact hit an enrollment trigger at 3am is a deliverability and experience problem. Check that send-time controls are configured on every outbound email action.

Data Quality and Enrichment

11. Duplicate detection rules are active and configured. Both HubSpot and Salesforce have native duplicate detection. It should be on. Check what the merge behavior is when duplicates are found.

12. Critical fields have a known fill rate. Pull a report: what percentage of contacts have a job title? Company size? Lifecycle stage? If any critical routing field is below 70% fill rate, you have an enrichment gap.

13. Enrichment is triggered on record creation, not just on demand. If enrichment only runs when someone manually clicks "enrich," the process still depends on manual follow-through.

14. Enrichment write-back rules are defined. Does enrichment overwrite fields reps have manually filled? It should not. Document and configure the write-back behavior per field.

15. Data decay policy exists. How old can a contact record be before it is re-enriched? Quarterly is a reasonable cadence for active contacts. If there is no policy, data quality is drifting.

Integration and Sync Health

16. Every active integration has an error log someone reviews. Zapier, Make, and native integrations all fail silently sometimes. Who reviews the error log? At what cadence? If the answer is "nobody," records are being lost.

17. Bidirectional syncs have conflict resolution rules defined. When both the CRM and the connected system update the same field, which one wins? If the answer is "I don't know," conflicting updates may be resolved inconsistently.

18. Webhook delivery is confirmed, not assumed. If your product or a third-party tool sends webhooks to your CRM, spot-check that recent events actually arrived. Webhooks fail silently when the receiving endpoint is down or rate-limited.

19. The integration between your CRM and your marketing automation tool is bidirectional. One-way syncs (marketing to CRM but not back) create split records and campaign attribution gaps. Verify that opt-outs, deal stage changes, and lifecycle stage updates flow both ways.

20. API rate limits are not being hit. Pull your CRM API usage report. If you are regularly near the rate limit, you have either a misconfigured integration running unnecessary calls or a sync architecture that needs rethinking.

Reporting and Governance

21. Your pipeline report and your finance forecast agree. Pull both and compare the top-line numbers. If they diverge by more than a rounding error, there is a data model mismatch between how deals are counted in each system.

22. Every custom field is used by at least one active workflow or report. Fields that exist but are not used in any automation or report are technical debt. Audit and clean them.

23. CRM admin access follows the principle of least privilege. Who can edit workflows, change field definitions, and modify pipeline stages? If the answer is "anyone," you will have accidental changes that break automation with no audit trail.

24. A change log or changelog process exists for the CRM. When someone adds a new workflow or modifies a trigger, is it documented? Without this, later investigations may lack the history needed to explain a failure.

25. You have run a simulated audit on a sample of recent deals. Take ten recent Closed Won and ten recent Closed Lost deals. Trace every automation that touched them. Did the right things happen at the right times? This is the real test of whether your CRM is working.

---

Run the audit before a major change to sales operations. Assign each finding an owner and verify the fix against recent deals.

Discuss CRM integration and automation findings with LinkIT.