Build audiences from your customer data, see who is in them, and use them in journeys.
Audiences are how you decide who a journey is for. You describe the group you want, Xenia works out who matches, and you get a counted audience with real numbers attached: how many people, what they spend, how engaged they are.
Open Audiences in the Orchestrate section.
Describe who you want to reach in plain language, and the AI writes the filter logic against your connected data. You can also build the filters by hand.
Building turns the definition into a live, counted audience with lifetime value and engagement stats.
Use the audience in a journey to decide who continues and who does not.
Xenia can only group customers it can see. Shopify, Klaviyo, and HubSpot are the sources the Audiences page works with directly. With none connected, the page asks you to connect one rather than offering templates that could not run. See Connectors.
The page offers six ready-made segments to start from, chosen from what your connected sources can actually support and mixed evenly across them. Picking one fills in the description for you, and you can edit it before building.
If you have HubSpot connected, the templates include company-based segments such as companies by city or by industry.
Choose Build with AI and describe the group in your own words, for example "customers who bought twice in the last 90 days but have not opened an email in 30". Xenia proposes the filters, explains who these people are, and shows how confident it is in the definition.
You can review and adjust every condition before building, so the AI is a starting point rather than a black box.
Each audience carries a goal, which tells the rest of the platform what you are trying to do with it:
Audiences start as drafts, become active when you are happy with them, and can be archived when you are done. The page filters by Active and Drafts so a half-finished idea does not get used by mistake.
Build audience runs the definition and produces a counted snapshot. Rebuilding later updates the same audience rather than creating a second one, so anything already pointing at it keeps working and you keep the history.
You can rebuild on demand with the refresh action on the card, or set a schedule so it keeps itself current.
An audience that has never been built shows as Not built yet. A built audience shows its member count and when it was last computed.
The number on the card reflects the last time the audience was built, not this second. Before making a decision on the size of a segment, check when it was last computed and refresh if it looks stale.
Open an audience to see whether it is worth targeting:
Engagement lift is the number to watch. A segment that is more responsive than your average customer is usually where a campaign pays off.
Compare audiences shows how much two audiences overlap: who is in both, who is unique to each, and an overall similarity score. Check this before launching two campaigns, so the same people are not messaged twice under different names.
From a built audience you can generate a lookalike, which uses the people in it as a model to find similar customers.
An audience is used as a check inside a journey rather than as a way to enrol people into one. The trigger decides who arrives, and the audience decides who keeps going.
In the journey builder, add an Audience Lookup node to check whether the person currently in the journey belongs to an audience. Follow it with a rule to split the path: your best customers get one offer, everyone else gets another.
The Use in journey action on an audience card builds a working example of this for you, which is the quickest way to see the pattern.