Unopened renewal notices, billing complaints and slow service tickets feed a weekly churn-risk score for every policyholder. When a score crosses the line, a retention task opens automatically for the assigned agent — complete with AI-drafted talking points for the save call.
Typical results
+6 pts
Annual retention
41%
At-risk clients saved
100%
Policies scored weekly
88%
Flagged before lapse
Figures reflect typical results for this kind of automation.
The challenge
The agency only discovered unhappy policyholders when the cancellation request arrived — by then the decision was made. The warning signs were all there, scattered across systems: renewal emails ignored, billing complaints logged, service tickets aging past SLA. No one connected them, so agents spent their energy on renewals that were never at risk while the quiet leavers slipped away.
Before and after
Annual retention rate
+7%At-risk clients contacted before lapse
4×Monthly cancellations
−37%The automation
We piped engagement signals from the email platform, billing system and helpdesk into one Make-managed data store, where every policyholder gets a weekly churn-risk score weighted by signal type and recency. When a score crosses the threshold, the scenario opens a retention task in the CRM for the assigned agent, and OpenAI drafts talking points from the client's actual history — what went wrong, what to offer, how to open the call.
How it works
- 01
Signals collected
Unopened renewal emails, billing complaints, aging service tickets and payment hiccups flow into one data store.
- 02
Risk scored
Every policyholder gets a weekly churn-risk score, weighted by signal type, severity and recency.
- 03
Task opened
A score crossing the threshold automatically opens a retention task for the client's assigned agent.
- 04
Talking points drafted
AI drafts the context: what happened, what to offer, and how to open the save conversation.
- 05
Outcome tracked
Saves, downgrades and losses are logged, feeding the dashboard and sharpening the scoring over time.
Starts from
- Renewal email signals
- Service & billing tickets
Delivers
- Retention task + talking points
- Assigned agent alert
- Risk dashboard
The impact
Agents now reach at-risk policyholders weeks before renewal instead of after the cancellation email, and 41% of flagged clients are saved. Annual retention climbed from 84% to 90% — six points that translate to hundreds of policies — while monthly cancellations fell from 67 to 42 without adding a single retention headcount.
Questions about this automation
Which signals feed the churn-risk score?
The score combines unopened renewal emails from the email platform, billing complaints and payment hiccups from the billing system, and service tickets aging past their SLA in the helpdesk. Each signal is weighted by type, severity and recency, and the agency can adjust the weights as it learns which signals best predict cancellations.
What does the agent receive when a policyholder is flagged?
A retention task opens in the CRM for the assigned agent, with AI-drafted talking points built from the client's actual history: what went wrong, what could be offered and how to open the call. The agent can edit or ignore the draft and decides what to say on the call.
How does the scoring improve over time?
Every flagged case is closed with an outcome, saved, downgraded or lost, and those outcomes feed the risk dashboard. Reviewing them shows which signals and thresholds produce useful alerts, so the weights can be tuned. It is a multi-step build with scheduled scoring, usually rolled out in stages.