We exported an Architect flow from our dev org and tried to import it into production. The import completed without errors, but the flow doesn’t work.
The flow references queue IDs, skill IDs, and Data Table IDs that exist in dev but have different UUIDs in prod. The import succeeded syntactically but the runtime references are all broken.
From a change management perspective, maintain a master UUID mapping document between all orgs.
Every queue, skill, Data Table, and integration has a different UUID in dev, staging, and prod. We maintain a spreadsheet that maps resource names to UUIDs across all 3 orgs. This is the bible for cross-org deployments.
This is the #1 reason to use CX as Code / Terraform instead of manual flow export/import.
Terraform resolves resource references by NAME at apply time. The same flow YAML works in both orgs because queue IDs are resolved dynamically via data sources. Manual export/import hardcodes UUIDs, which are org-specific.
If you must use manual import, search-and-replace the UUIDs in the flow YAML before importing.
import json
def remap_flow(flow_yaml, id_map):
content = json.dumps(flow_yaml)
for old_id, new_id in id_map.items():
content = content.replace(old_id, new_id)
return json.loads(content)
id_map = {'dev-queue-uuid': 'prod-queue-uuid',
'dev-skill-uuid': 'prod-skill-uuid'}