Create, update, and publish Data 360 segments by describing the audience to an AI client connected to the Data 360 MCP server. The server builds the filters, counts the members, and publishes the segment the same way the segment canvas does.
Where: This change applies to Data 360 in Enterprise, Performance, Unlimited, and Developer editions that include segmentation.
When: This feature is available starting in October 2026.
Why: You can build an audience without opening the segment canvas or writing filter criteria. The server chooses the segment-on entity, resolves fields and relationships, and asks a clarifying question when the request is ambiguous. For example, if you say "customers," it can ask whether you mean individuals or accounts.
How: Connect an AI client to the Data 360 MCP server. Your org needs a data space and a mapped, populated data model. Identity resolution is recommended when the audience uses unified profiles. Then describe what you want. For example:
- Create and publish: "Create a segment of individuals in the US who bought at least once in the last 30 days, then publish it."
- Filter on a calculated insight: "Segment individuals whose lifetime value is greater than 10,000."
- Combine segments: "Create a segment of Loyalty Program Members and exclude the Newsletter Subscriber segment."
- Update an existing segment: "In the Recent US Buyers segment, also include Canada."
- Set a refresh schedule: "Publish Recent US Buyers every 24 hours."
- Check the result: "How many people are in Recent US Buyers?"
The server creates a standard segment unless you name another type, such as a dynamic, real-time, or waterfall segment. You can also list, rename, change the lookback period, and deactivate segments. You can't change a segment's API name or type after it is created.