Data platform export

The data platform connector sends content-free facts from DialSplice into your own warehouse: call aggregates, work-state transitions, appointment outcomes, and the analytics figures behind speed to lead and show rate. The export runs on your architecture, on Microsoft Fabric / OneLake, Microsoft Azure, AWS, or Google Cloud, under your governance. It never carries contact identity, message bodies, recordings, or provider tokens.

What is exported

• Call aggregates per location and view.
• Work-state transitions with timestamps.
• Appointment outcomes: kept, cancelled, no-show.
• Analytics facts: speed to lead, connection, schedule, and show rate.
• Export batch manifests with hashes.

DialSplice reads only connection health and schema version from your platform. It writes nothing outside the allowlisted attributes.

Steps to set up an export

1. Ask your onboarding contact to enable the data platform connector for your cloud. Credentials are configured server-side.
2. Review the export schema. It is part of the location configuration, so it appears in a revision with a content hash.
3. Approve and activate the revision. Readiness checks confirm the mapping is valid before activation.
4. Confirm the first batch lands. Each batch is a ledger event with a hash and a manifest.
5. Point your own reporting at the tables. The rows match what the DialSplice analytics view shows.

What you’ll see

• The connector card shows “Data plane”, “Server-side credentials”, and the status: Supported for Azure, AWS, and Google Cloud; Connecting for Fabric / OneLake.
• The ledger records each export batch. A failed batch is held for reconciliation, not silently retried.
• The analytics page note reads “read back from your system of record” for show rate, so the same figures are in your warehouse.

Changing the mapping

Mapping changes are revisions. Preview the schema, approve, activate, and roll back if a downstream model breaks. Old batches keep the schema version they were written with.

If something looks wrong

• Batches stop arriving. Check connector status and the reconciliation holds on the manager desk.
• A column you expected is missing. It is probably outside the allowlist. Identity and content are never exported by design.
• Numbers differ from your CRM report. DialSplice counts from the ledger, not self-report. Compare against the analytics view first.

Related

• Connector status and credentials
• Exporting the call log
• Reading the audit ledger