data360-segment

This skill covers the Segment phase of Salesforce Data Cloud, focused on audience and insight work: creating and publishing segments, managing calculated insights, checking member counts, and troubleshooting audience SQL. It solves the problem of building and validating Data Cloud audiences reliably by giving the assistant clear task ownership, a repeatable workflow, and guidance on the quirks that make segment work fail silently — such as asynchronous publish steps and segment SQL not being the same as CRM SOQL.

It operates through the external community sf data360 CLI plugin against a Data Cloud-enabled org. The documented workflow runs a shared readiness classifier from the companion data360-orchestrate skill, inspects existing segments and calculated insights, creates them from reusable JSON definitions (using an explicit API version when creation behavior is unstable on newer defaults), publishes or runs them explicitly, and then verifies success with member counts or SQL queries rather than assuming completion. It captures high-signal gotchas: segment creation sometimes requiring api-version 64.0, segment members returning opaque IDs (so SQL joins are preferred for readable details), publish/run kicking off asynchronous work, and an empty list usually meaning unconfigured rather than unavailable. It delegates DMO/mapping/identity-resolution work to data360-harmonize, activation to data360-activate, and read-only SQL/search-index queries to data360-query. These are standard, org-scoped administrative operations on the authenticated org.

The target users are Salesforce Data Cloud administrators, marketers, and consultants building audiences. Use cases include creating a high-value customer segment, diagnosing why a segment returns zero members, running and verifying a calculated insight, and getting member counts.