This recipe describes a practical way to reveal churn risk using the data you already have.
Churn risk often appears first in behavior and conversations rather than in high-level KPIs. Complaints, stalled onboarding, or reduced usage can all be early warnings. A churn signal assistant scans existing data and turns it into short narratives and labels that teams can act on.
When using Unless, you can simply switch on churn risk monitoring in the Retention Lifecycle Stage. This will make sure that conversations are labeled if end users are at the risk of churning, according to default rules or rules that you may set yourself. This can also lead to a Task in the Task Manager, or even to notifications for your account team.
In the front-end, for the end user, it works differently. There, you can define an AI Skill that triggers based on a certain Audience, or on certain questions. This skill may pop up as soon as somebody signals that they are about to leave as a customer, for example when they ask for an export of all their data. The skill may then inquire if they would be interested in a discount or additional feature, or whatever solution is relevant to your use case.
To keep explanations honest and grounded, limit the skill to flat Topics such as Health-signals, Retention-playbooks, and Customer-outcomes. These Topics can contain your own definitions of what “risky” behavior looks like and recommended responses, so the AI mirrors your success practices rather than inventing its own model.
So, in short, when certain churn signals appear, the skill can automatically create tasks for the owning CSM or account manager. It can also send notifications summarizing new at-risk accounts into a team channel or dashboard each week. This makes risk visible without requiring people to manually scan reports.
For accounts at risk but still active in the portal, in-app notifications can be used carefully. Based on segmentation by Audience, you might show targeted educational content or invite users to book a review call. These should focus on value and support, not pressure, and must respect any regulatory constraints.
Conclusion
A churn signal and health score surfacing assistant turns scattered traces of risk into clear signals for your teams. With an AI Skill, flat Topics that encode your view of health, and structured tasks and notifications, it helps you prioritize retention work and respond before it is too late.
