A caller corrects the service address, and the GoHighLevel call summary contains the new value. The custom field used by dispatch still contains the old address. The conversation captured the correction, but the operational write did not.
Custom-field mapping needs a source, validation rule, and update policy for each field. A summary is not a substitute for that contract.
Map confirmed values to the intended fields
For a fictional service intake, distinguish the contact's usual address from the address for this specific request. Updating the wrong field can alter future work unrelated to the call.
| Mapping decision | Question |
|---|---|
| Source value | Was it caller-confirmed or inferred? |
| Destination field | Does it represent the contact or this enquiry? |
| Validation | What format and allowed values apply? |
| Update permission | May this workflow change it? |
| Missing value | Preserve, clear, or request review? |
Verify current custom-field capabilities and identifiers in the client's account against HighLevel's documentation. Do not map by a display label that may be duplicated or renamed.
Preserve correction order through delayed work
Keep the latest confirmed request version with its pending writes. If the caller corrects a date twice, an older queued update should not overwrite the final value.
Do not automatically clear existing data because a later payload omitted a field. Missing, deliberately cleared, and unknown are different states that need an agreed policy.
Confirm the result where staff actually use it
After the write, inspect the record and any downstream task or workflow. A successful API response does not prove that a later automation did not overwrite the value.
If the source and destination represent different concepts, document the transformation. A preferred contact window should not become a confirmed appointment time merely because both are stored as text.
The appointment-outcome guide describes the same modelling issue in another CRM context.
Test edits, omissions, and conflicts
Use fictional calls with a corrected address, corrected date, empty optional field, explicit clearing request, and two queued versions arriving out of order.
Inspect both the field values and generated messages. Keeping the CRM correct while sending stale details to staff is still a failed workflow.
For the first pilot, map only fields with a clear operational purpose. A large custom-field inventory increases maintenance without guaranteeing better intake.
Custom-field questions
Is the call summary enough evidence that a field changed?
No. Inspect the actual destination value and its downstream use.
Should omitted fields always be cleared?
Use an explicit update policy rather than assuming omission means deletion.
What belongs in the handover?
Provide a mapping table, identifiers, source rules, and correction tests in the integration pack.