The input JSON paths In the DataRaptor Load must match the JSON from the OmniScript, otherwise the data will not be mapped correctly to the Salesforce fields. ContextId must be spelled correctly in the merge code that passes the Recordld to the DataRaptor, otherwise the DataRaptor will not be able to identify which record to update. The fields that are not updated are not necessarily read only in Salesforce, as they could be editable by other means. In the DataRaptor Load, in the Preview tab, the Record is from the wrong record type has no effect on updating data in Salesforce, as it is only used for testing purposes.