Skip to main content

Smart Reports: Churn Risk Report

How the Churn Risk report shows which calls signal a customer may leave, how that's changing, and which themes are driving it.

C
Written by Cameron Parry

What is the Churn Risk report?

The Churn Risk report shows which customer calls signal a risk of the customer leaving, how that's changing over time, and which themes are driving it. It's built on Kapiche's Churn Risk AI enrichment, which reads each call and classifies how likely the customer is to leave based on what they said.

Each call is given one of three levels:

  • High: explicit churn intent, such as cancellation threats, requests to close the account, severe dissatisfaction, or threats of legal or regulatory action.

  • Potential: warning signs without explicit churn intent, such as repeated complaints, expressed frustration, declining engagement, or mentions of competitors.

  • Low: no meaningful sign of churn.

Calls at High or Potential risk count as at risk. Low risk calls are the comparison group.

Getting a Churn Risk report

The Churn Risk report is set up by the Kapiche team, and it needs the Churn Risk AI enrichment running on your call data. If you'd like one, ask your Customer Success Manager. Once it's set up, you can view, filter, share and schedule it like any other Smart Report. By default it covers the last month.

What's in the report

Churn Risk Summary

Tiles show the number of High risk and Potential risk calls in the period, with the change since the previous period.

Churn Drivers

A table of the themes raised by calls at risk of churn, and how much more often they come up than in other calls:

  • % of at-risk calls: the share of at-risk calls that mention the theme.

  • Times more likely: how many times more likely an at-risk call is to mention the theme than a low-risk call. Only at-risk means the theme appears only in at-risk calls.

  • p: how likely it is that the difference is down to chance. Rows are greyed out when there are too few calls to rule chance out (p > 0.05), so treat those with caution.

Each value shows its change since the previous period. Click a column header to sort, then click a theme to open the calls behind it, already filtered to that theme and the at-risk levels.

Record Table

A table of the individual calls, which you can filter by theme. Open a call to read the transcript along with its Churn Risk level and the Churn Risk Reason, the AI's explanation of why the call was flagged.

Tips

  • Start with Churn Drivers sorted by Times more likely to find the issues most specific to at-risk customers. Then switch to % of at-risk calls to see which of them affect the most people.

  • A theme can be strongly linked to churn but rare. Check both columns before prioritising.

  • Read the Churn Risk Reason on a few calls for each top driver. It's the quickest way to understand what customers are actually unhappy about.

Related articles

Did this answer your question?