The backbone
The self-healing automation backbone
Nothing reconciles itself unless something is watching.
Leads arrived from 8 portals and ad sources, each in its own format, got mistagged or dropped, and the CRM no longer matched what was happening in the field. Month-end billing was a manual reconciliation. Nothing reconciled itself, so every gap needed a human to notice it, and at this volume humans stopped noticing.
8 lead sources, 8 formats, 1 CRM
An automation layer on n8n Cloud that sits between GoHighLevel and the rest of the stack and continuously reconciles all of it without anyone watching. The real work is the audit-and-repair layer, not the connections.
re-runs are safe: dedupe guards on every write
Base44 to Google Drive sales sync
54 nodesMirrors every sale from the client's property app into a structured Drive hierarchy with duplicate-proof folder logic, invariant guards, and a writeback loop. The largest workflow in the suite.
Asana to Xero monthly billing
41 nodesTurns tracked time into validated, deduped Xero invoices, replacing a manual month-end reconciliation. A dedupe guard prevents double-billing on re-runs.
EOI to CRM sync
36 nodesSyncs expression-of-interest submissions with a sparse, non-destructive patch that fills only blank fields and never overwrites agent-entered data, with a protected-stage guard.
Appointment state machine
33 nodesReal-time bookings and cancellations with stage rollback and inspection-tag lifecycle handling, so the CRM and the reporting layer always match what happened.
Self-healing tagging and reconciliation
~8 workflowsAuto-creates missing pipeline opportunities, audits tags, and recreates dropped leads across 5 pipelines on a schedule. Drift gets detected and repaired automatically.
Multi-portal lead-parser fleet
8 workflows8 portal and ad-source parsers converge on a single enrichment-and-routing pattern with a redundancy catch-net, so every lead lands the same way and none drops silently.
90+
workflows built, 64 live across 10 integrated platforms
HandoverEvery workflow is documented, failures alert to Slack, and each automation has an off switch the team can reach without me.
drift gets detected, then repaired. no ticket raised
Evidence · fig. 01–06